@sdelsad/commodity-desk-daily 1.0.19 → 1.0.21

package/covered.md

@@ -8,3 +8,4 @@ Running log. Read before writing a new episode: avoid repeating material, and on
 - **Ep 4** (Thu) — *The Physical Chain, End to End*: Incoterms as risk allocation (FOB/CFR/CIF, risk passes at loading, cost vs risk separate, who charters/insures); execution clock laycan-nomination-NOR-laytime-demurrage/despatch; worked example 60kt FOB Santos beans at ~434 USD/t = 26M cargo, 3 days over at 24k/day = 72k vs 660k margin (11%), interest 4.3k/day; statement of facts and cascading demurrage claims; laycan miss = cancellation into a 40c rally; documents: draft survey, certificate final at load, bill of lading as title, backdating = fraud; execution desk as profit centre; OPS/TRADER dialogue on NOR and turn time. Vocab: Incoterms, CFR, CIF, charter party, nomination, NOR, laytime, weather working day, despatch, statement of facts, draft survey, bill of lading, cancelling date. Pulse: WASDE aftermath - corn yield cut to 180.7 (trade 182.5, prior 183), new-crop ending stocks 1.653bn vs 1.79 July, Dec corn +20.25c to 4.8075 two-week high; beans production +44M above July yet Nov +14.5c to 11.8325 on crush +30M (trade whole sheet, not one row); Chi wheat +22.5c to 6.5275, KC +21.5c to 7.2075; GEO escalation: Tue talks rumour died overnight, Ukraine struck Novorossiysk idling Demetra (8.5Mt) + NKHP (7.1Mt) grain terminals ~15.5Mt/yr, Russian Aug exports est 3.0-3.4Mt, Turkey two-corridor proposal, vessels-on-demurrage-clock bridge into lesson
 - **Ep 5** (Fri) — *Wheat: The Map and the Screens*: Wheat classes and specs (SRW ~10 Chicago, HRW 11-12.5 KC, HRS 13.5+ Minneapolis, durum, Black Sea milling 11.5-12.5); protein, test weight and falling number as the real price, low falling number demotes milling to feed at ~40 USD/t. Four exchanges for one grain: Chicago and KC 5000 bu in c/bu, Minneapolis HRS, Matif EU milling 50 t lots in EUR/t delivered Rouen-Dunkirk; tick symmetry 12.50 dollars vs 12.50 euros; 60kt = 440 Chicago lots vs 1200 Matif lots. KC over Chicago 68c/bu = 26 USD/t as the protein spread and an export-bid signal. Black Sea has no futures - daily price assessments, why an assessment cannot be bought sold or hedged. Cross-hedge worked example: 60kt Russian 12.5 FOB at 224 hedged with 1200 Matif lots, Europe +10 EUR/t = -692k against physical +4 USD/t = +240k, net -452k = 7.5 USD/t slippage; cross-hedge protects against the world moving not your own market; correlation highest on quiet days; EUR/USD exposure created by the hedge itself (~13-14m EUR). MILL/SELLER dialogue on protein, falling number, test weight and the 9-dollar spec spread. Pulse: Thu 13 Aug give-back - Dec corn 4.7775 -0.6 percent, Nov beans 11.8175 flat, Chi Sep wheat 6.5125 -0.2 percent, KC Sep 7.2075 Wed settle; China bought new-crop US beans three days running totalling 505,000 t; GEO escalation - Russia struck Izmail on the Danube, Ukraine's fallback after deepwater loadings ~zero since 22 July, Ukrainian early-Aug shipments -76 percent y/y, wheat export forecast 8.3 Mt, USDA cut Russia+Ukraine exports 2.5 Mt, yet Chicago finished the week unchanged because US sales were only 255,900 t (-14 percent w/w) and the US share of world trade was cut to 9.9 from 10.9 percent - flow substitution needs a buyer who actually switches origin, and they call France, Argentina and Australia.
 - **Ep 6** (Mon) — *Corn, Crop Calendars and Weather Risk*: Corn as a demand story (feed ~2/5, ethanol grind and its margin switch, exports 3.275bn bu, stepped demand curve); corn-wheat feed substitution priced both ways - Dec corn 477.5 = 188 USD/t vs Dec SRW 679 = 249.5 USD/t, 4 percent feeding credit gives a 195 USD/t switch level, 54.50 USD/t gap = 148 c/bu, wheat would need 531; 654k a month on a 20kt mill at 60 percent inclusion; reverse ceiling corn at 240 USD/t = 609 c/bu; BROKER/FEEDER dialogue quoting flat-to-corn rather than a wheat price. Crop calendar table US/Ukraine/Brazil full-season/safrinha/Argentina, US and Ukraine share a hemisphere so not diversified, safrinha is 3/4 of Brazilian corn and its risk is the soybean harvest date in front of it (wet October to May pollination in the dry season). Anatomy of a weather premium: price of a distribution vs trend yield, builds 10-14 days before the window, decays on the calendar not the forecast; Aug WASDE case - yield cut 183 to 180.7 removed 204m bu on 88.6m harvested acres but 2.8m acres found lifted production to 16.013bn, second largest ever; planted vs harvested acres as two denominators, ~8m acres never cut for grain. Pulse: Fri 14 Aug closes higher across the board with wheat up 4 percent on the week (Sep corn 459 +11, Sep beans 1177.75 +11.75, Sep meal 310.20, Sep oil 69.44, Sep Chi wheat 674.75 +22, Dec SRW 679, Dec KC 747.25, Dec corn 477.5, Matif spot 228.25 EUR); GEO escalation - all three Novorossiysk grain terminals suspended by Ukrainian drone strikes, Russian August loadings ~2.5 Mt = under half the five-year pace and weakest August since 2016/17, Ukraine MTD 201.7 kt -76 percent y/y, deepwater corridor shut since 22 July, Russia rejected partial ceasefire for civilian shipping - flow substitution moved from threat to actual buying, which is why this week the price moved and last week it did not
+- **Ep 7** (Tue) — *WASDE and Building a Balance Sheet*: How a grain balance sheet is built line by line, and why ending stocks — the line nobody measures — moves about ten times faster than the crop itself. Plus feed and residual, the line that hides the sins, and why two competent analysts agree on supply and fight about demand.

package/ep07.md

@@ -0,0 +1,280 @@
+# Market pulse
+
+**Monday belonged to soybeans, and they took it on demand rather than supply.**
+
+| Commodity | Contract | Price | Change |
+|---|---|---|---|
+| Corn | Sep (CBOT) | 465 c/bu | +6¢ |
+| Corn | Dec (CBOT) | 489½ c/bu | +6¼¢ |
+| Soybeans | Sep (CBOT) | 1201 c/bu | +23¼¢ |
+| Soybeans | Nov (CBOT) | 1216 c/bu | +23½¢ |
+| Wheat HRW | Sep (KC) | 758¾ c/bu | +4½¢ |
+
+Monday's session belonged to the soybean complex, which added better than 1¾%. Chicago soft red wheat was the exception, giving back part of a week in which it had gained 35 cents. Kansas City held its bid.
+
+Two things did the buying. China took another 136,000 t of US soybeans. And the crop split in two: the northwestern belt is dry, the eastern belt has had too much rain. Monday afternoon's condition ratings had corn at 60% good to excellent and soybeans at 61%, each a point lower on the week, with soybeans seven points below where they stood a year ago. Corn is 76% at dough and 4% mature, running ahead of the five-year average. The bean crop is 85% setting pods.
+
+**The geopolitical read: the buyer's clock.** Chinese purchases of new-crop US soybeans have been running far ahead of anything that has actually sailed. Vessel data through late July put US-to-China departures at roughly 11,000 t/day on a seven-day average — close to nothing against commitments running into the millions of tonnes. That gap is an instrument, not an accident. A commitment is a promise on a balance sheet. A loading is a fact on a vessel. The buyer decides when one becomes the other, and that timing is worth more than the price paid. It also lands on exactly one line of today's subject.
+
+```chart
+{"type":"line","mode":"index","unit":"index, Tue 11 Aug = 100",
+ "title":"Corn led the week, beans closed on Monday",
+ "x":["Tue 11","Wed 12","Thu 13","Fri 14","Mon 17"],
+ "series":[{"name":"Corn Dec (CBOT)","values":[460.50,480.75,477.75,483.25,489.50]},
+           {"name":"Soybeans Nov (CBOT)","values":[1168.75,1183.25,1181.75,1192.50,1216.00]}],
+ "caption":"Both jumped on the 12 August WASDE and neither gave it back. Corn has led since the report, but Monday belonged to beans, which added 2.0% against corn's 1.3% on Chinese buying and a dry northwestern belt.",
+ "source":"CBOT settlements, 11–17 August 2026, from daily market wraps."}
+```
+
+# Key takeaways
+
+- A balance sheet is not a forecast. It is an accounting identity that has to close, and ending stocks is the line that closes it.
+- Production is roughly ten times the size of US corn ending stocks, so a 1% error in the crop is a 10% error in the carryout. Work that ratio out for every commodity you trade — it is different for each one.
+- Stocks-to-use maps to price on a curve, not a line. The same two-bushel yield move is worth almost nothing at 15% and an enormous amount at 9%.
+- Feed and residual is a residual inside a residual. When it moves, it is telling you about livestock or telling you the crop was never the size they said it was.
+- Analysts converge on supply and diverge on demand, because supply is surveyed and demand is inferred. That is why pre-report positions live in spreads.
+- The market does not price the yield. It prices the carryout, and area, demand and carry-in can each pay for a bad yield.
+
+# Vocabulary
+
+| Term | Meaning |
+|---|---|
+| **Balance sheet (S&D)** | The one-page supply and demand statement for a crop and a marketing year, which must close |
+| **Marketing year** | The crop's own accounting year — September to August for US corn and soybeans, June to May for US wheat |
+| **Carryout** | Ending stocks, spoken as a single word on the desk |
+| **Total supply** | Carry-in plus production plus imports, the top block of the sheet |
+| **Total use** | Domestic use plus exports, the bottom block |
+| **Feed and residual** | The inferred demand line that absorbs both livestock feeding and every measurement error in the sheet |
+| **New crop / old crop** | The marketing year about to begin versus the one ending, priced by different contract months |
+| **Trade average** | The mean of analysts' pre-report estimates, published in advance, and therefore what is already in the price |
+| **Whisper number** | The expectation the market actually trades, which can sit away from the published trade average |
+| **Crop Production** | The NASS report published alongside WASDE, carrying the survey-based yield and area |
+| **Grain Stocks** | The quarterly survey of physical inventories, from which feed and residual is backed out |
+| **NASS** | USDA's National Agricultural Statistics Service, the body that runs the surveys behind the numbers |
+| **Implied disappearance** | Use derived by subtraction rather than measurement, the technique behind the residual lines |
+
+# Quiz
+
+**Q1.** Build one. A desk runs its own 2026/27 US corn sheet with carry-in of 1.945 bn bu, imports of 25 m bu and harvested area of 88.6 m acres, but uses its own yield of 179.0 bu/ac. It carries domestic use at 13.100 bn and exports at 3.350 bn. Compute production, total supply, ending stocks and stocks-to-use. Then compare the percentage gap between the desk's carryout and USDA's 1.653 bn with the percentage gap between the two yields, and explain the difference.
+
+**Q2.** The quarterly Grain Stocks survey lands 90 m bu below what the trade had modelled, with no change to production or exports. Which line absorbs it, what are the two competing stories for why, and how would you use the futures *curve* rather than the flat price to work out which story the market believes?
+
+**Q3.** Going into 12 August the trade carried new-crop corn carryout at roughly 1.725 bn bu. USDA printed 1.653 bn. December corn rose 20¼¢ on the day and closed Monday 17 August at 489½, above its report-day close. Explain why "the report was bullish and the market went up" is a lazy reading of those three sessions, and state precisely what a trader had to be right about to still be paid on Monday.
+
+**Q4.** Ep 6 said corn's demand curve has steps in it. USDA carries corn exports at 3.275 bn bu and domestic use at 13.055 bn. Suppose a rally takes December corn from 489½ to 560 c/bu and nothing else changes. Name the two demand lines that respond first and the direction each moves. Then explain why a balance sheet that leaves demand untouched after a 70-cent rally is internally inconsistent.
+
+**Q5.** Ep 6 taught that a weather premium decays on the calendar rather than on the forecast. It is 17 August. Corn is 76% at dough and 4% mature. Soybeans are 85% setting pods and rated seven points below last year. A trader wants to be long the weather. What is actually left to be long of in each crop, which would you rather own, and how does Monday's tape support the answer?
+
+**Q6.** Ep 4 established that risk passes at loading under FOB, CFR and CIF alike. You have sold 60,000 t of beans CFR Qingdao. The vessel is three days late arriving after a typhoon diversion, and your buyer's letter of credit expires in four days. Separate the voyage risk from the documentary risk, say who carries each, and identify which of the two actually threatens your P&L.
+
+**Q7.** Ep 4 priced three days of demurrage at $24,000/day as 11% of a $660k margin. Do it in the other unit. A 60,000 t soybean cargo is valued off Monday's November board of 1216 c/bu, and demurrage runs three days at $26,500/day. Convert the cargo to bushels and dollars, then express the demurrage in cents per bushel and as a percentage of cargo value. Why do desks insist on the cents-per-bushel version?
+
+**Q8 — Conversion drill.** Over the weekend an eastern Corn Belt station recorded 82 mm of rain, while a western Kansas station is forecast 0.35 inch for the week ahead. Convert each figure into the other scale. Then say which of the two matters more for the balance sheet that moved this week, and name the line it lands on.
+
+# SOLUTIONS (spoilers)
+
+**A1.** The arithmetic first.
+
+| Line | Desk | USDA (Aug) |
+|---|---|---|
+| Carry-in | 1,945 | 1,945 |
+| Production | 15,859 | 16,013 |
+| Imports | 25 | 25 |
+| **Total supply** | **17,829** | **17,983** |
+| Domestic use | 13,100 | 13,055 |
+| Exports | 3,350 | 3,275 |
+| **Total use** | **16,450** | **16,330** |
+| **Ending stocks** | **1,379** | **1,653** |
+| Stocks-to-use | 8.4% | 10.1% |
+
+All figures in million bushels. Production is 88.6 × 179.0 = 15,859. (USDA's published 16,013 is about 3 m bu above 88.6 × 180.7 — rounding in the acreage and yield they print. Worth noticing the first time you try to rebuild a published sheet and cannot make it tie.)
+
+Now the comparison the question is really asking for. The yields differ by 1.7 bu, or **0.94%**. The carryouts differ by 274 m bu, or **16.6%**. That is a factor of roughly eighteen, not ten, and the extra leverage comes from the demand side: the desk is carrying 120 m bu more use than USDA. Decomposed, the 274 is 154 m of yield-and-rounding and 120 m of demand. The trap the question sets is the assumption that a carryout gap is a crop-view gap. Most of the time it is half a crop view and half a demand view, and only one of those two halves gets discussed on television.
+
+**A2.** **Feed and residual absorbs it,** and ending stocks falls 90 m bu — from 1.653 to 1.563 bn, taking stocks-to-use from 10.1% to 9.6%. Nothing else in the sheet is allowed to move, because feed and residual is the line derived by subtraction.
+
+The two stories:
+
+1. **Real demand.** Livestock ate more than modelled — bigger herds, cheaper corn, better rations. This is a genuine tightening and it should persist into the next quarter.
+2. **Phantom supply.** The crop was never as big as printed. The stocks survey is measuring physical inventory, so if it comes in light, either use was higher or production was overstated. A production overstatement gets revealed slowly, through the residual, quarter by quarter.
+
+The curve tells you which one the market believes. If the trade takes it as real, present demand is tighter *now*: the nearby month bids relative to the deferred, spreads firm, and a carry market flattens or inverts. That is bull spreading, and it is a statement about physical availability. If the trade takes it as a statistical artefact likely to be revised away, flat price wobbles on the headline and the spreads do not move — the curve is saying nobody is short of corn today. **Flat price reacts to news. Spreads react to grain.** A move in one without the other is the market telling you how much it believes.
+
+**A3.** The surprise was **72 m bu**, not the 159 m bu that the 1.8 bu/ac yield miss implies on 88.6 m acres. Half of it was absorbed on the way down the page by higher area and higher exports. In stocks-to-use terms the trade went in at 1,725 ÷ 16,330 ≈ 10.6% and got 10.1% — about half a point of tightening. That is what the 20¼¢ paid for, and it is a reasonable price for it.
+
+What happened afterwards was a different trade. Corn did not hold its gain and then add another six cents on Monday because of the report. It did so because Black Sea export capacity stayed impaired, because the eastern belt turned too wet, and because the whole complex was being pulled up by soybeans on Chinese demand. None of that was in the WASDE.
+
+So a trader long into the print was paid twice, for two separate reasons, and needed to be right about both to still be there on Monday. Being right about the report bought about a day. Staying long required an independent view on export capacity and August weather. The lazy reading — "bullish report, market up" — conflates an event that resolved in ninety seconds with a trend that has been running for two weeks. The discipline is to book the event P&L, then re-underwrite the position on the trend as a fresh decision.
+
+**A4.** **Exports fall, and the ethanol grind inside domestic use falls.** Both are price-elastic on a timescale of weeks.
+
+- **Exports** are the fastest. US corn competes with Brazilian, Argentine and Ukrainian corn at destination. A 70¢ rally is about $27.50/t; the buyer simply calls another origin. This is ep 5's flow substitution, running on the demand line instead of the supply line.
+- **Ethanol grind** is a margin, not a preference. The plant buys corn and sells ethanol and distillers grains. Corn up 70¢ with ethanol unchanged compresses the crush, and marginal plants slow down. Demand stops appearing at the bid.
+
+Feed is the third lever but the slowest and, right now, the furthest away: ep 6 put Chicago wheat about $54/t above its feed-switch level, so the substitution bid is nowhere near being triggered.
+
+The inconsistency is this. A balance sheet is a set of **quantities at an assumed price**. If you tighten stocks and let price rise, you must also let the demand lines respond, or you are counting the tightening twice. That feedback is why carryout estimates converge rather than diverge: tighter stocks lift price, higher price rations demand, rationed demand rebuilds stocks. Rationing is not a metaphor — it is the arithmetic by which the sheet closes at a higher price. The analyst's edge sits almost entirely in how fast and how far they think that response runs.
+
+**A5.** For **corn**, very little is left. Pollination determines ear count and it is long finished; 76% dough and 4% mature means kernel number is set and the crop is filling. What remains is test weight, late-season disease, and an early frost risk that is still weeks away and low-probability. The weather premium has largely decayed, exactly as ep 6 described, and it decayed whether or not the forecast improved.
+
+For **soybeans**, the crop is genuinely still open. Beans set pods and fill seed through August, so August rainfall and August heat still move the yield. With 85% setting pods, a rating of 61% good to excellent and seven points of deterioration against last year, there is a real distribution left to price.
+
+You would rather own the beans — and Monday's tape is the evidence: beans +23½¢ against corn +6¼¢ on the same weather map. The sharper version of the answer is that you would rather own *optionality* on beans than futures, because the same decay clock that has already emptied the corn premium is running on the bean premium too, and it accelerates from early September. A long call spread pays for the distribution without paying carry to the calendar.
+
+**A6.** Two different risks, and the question is whether you can tell them apart under pressure.
+
+- **Voyage risk** passed to the buyer at the ship's rail in Santos. A typhoon diversion mid-ocean is the buyer's risk of loss, not yours. Under CFR you contracted the freight, so a late vessel is your dispute with the carrier under the charter party — but it does not move cargo risk back onto your book.
+- **Documentary risk** is entirely yours, and it is the live one. A letter of credit is a bank's undertaking to pay against **conforming documents presented before expiry**. Miss the expiry and the undertaking lapses. You are then an unsecured creditor of the buyer, holding a cargo already at their risk, in a market where they may prefer to renegotiate.
+
+**The documentary risk is what threatens the P&L.** Nothing has happened to the beans. The practical moves are to request an L/C amendment extending expiry and the latest shipment date before it lapses rather than after, to present documents early if the bill of lading and quality certificate are already in hand, or to ship documents against a letter of indemnity if originals are chasing the vessel. Ep 4's point restated: the cargo is fine, and the paper is what pays.
+
+**A7.** Convert first. 60,000 t × 36.744 bu/t = **2,204,640 bu**. At 1216 c/bu the cargo is worth **$26.81 m**.
+
+| Line | Value |
+|---|---|
+| Cargo | 2,204,640 bu |
+| Cargo value at 1216 c/bu | $26,808,422 |
+| Demurrage, 3 days × $26,500 | $79,500 |
+| Demurrage in c/bu | 3.6 c/bu |
+| Demurrage as % of cargo value | 0.30% |
+
+Desks quote it in cents per bushel because that is **the same unit as the margin**. A cargo bought at "November minus 20" and hit with 3.6 cents of demurrage was really bought at minus 23.6. In dollars, $79,500 against $26.8 m looks like a rounding error and gets waved through. In cents per bushel it sits next to the differential that the entire trade was argued over, and 3.6 cents is a fifth of the twenty. Execution costs only become visible when they are expressed in the unit the trader negotiates in — which is why the execution desk reports in cents, not in invoices.
+
+**A8 — Conversion drill answer.** 82 mm ÷ 25.4 = **3.23 inches**. 0.35 inch × 25.4 = **8.9 mm**. The fast method gets you close enough to trade on: 82 ÷ 100 × 4 = 3.28, and 0.35 × 100 ÷ 4 = 8.75.
+
+The **82 mm** is the number that matters this week. Excessive rain across the eastern Corn Belt in mid-August, with corn 76% at dough and beans 85% setting pods, brings lodging and disease pressure into crops that are filling, and it is behind the one-point slip in both condition ratings on Monday. It lands on the **yield term of production**, on the supply side of the 2026/27 sheet.
+
+The 8.9 mm in western Kansas is a soil-moisture story for hard red winter seeding in September. It reaches a balance sheet too — but the 2027/28 one, through planted area and abandonment. Same week, two rainfall figures, two different marketing years. Keeping them apart is most of the skill.
+
+# Written edition
+
+## The object every grain desk argues about
+
+A balance sheet is one page. Supply on top, demand underneath, and what is left over at the bottom. It is not a forecast and it is not a model. It is an accounting identity, and it has to close.
+
+Supply is three lines.
+
+| Line | What it is |
+|---|---|
+| Carry-in | What was still in store when the previous marketing year ended |
+| Production | Harvested acres × yield per harvested acre |
+| Imports | For US corn, a rounding error |
+
+Two things in that table are quietly load-bearing. A **marketing year** is the crop's own calendar, not the accountant's: US corn and soybeans run September to August, US wheat June to May. And production uses **harvested** acres, not planted. Around 8 million planted US corn acres never get cut for grain in a normal year.
+
+Demand is four lines: feed and residual, ethanol, food/seed/industrial, and exports. Then ending stocks, which is total supply minus total use.
+
+Ending stocks is not measured. **It is what is left.**
+
+## Building the real one
+
+The August numbers, in million bushels.
+
+| | 2026/27 US corn |
+|---|---|
+| Carry-in | 1,945 |
+| Production (88.6 m ac × 180.7 bu/ac) | 16,013 |
+| Imports | 25 |
+| **Total supply** | **17,983** |
+| Domestic use | 13,055 |
+| Exports | 3,275 |
+| **Total use** | **16,330** |
+| **Ending stocks** | **1,653** |
+| **Stocks-to-use** | **10.1%** |
+
+```chart
+{"type":"waterfall","unit":"million bushels",
+ "title":"US corn 2026/27, supply down to carryout",
+ "caption":"Ending stocks is the difference between two numbers that are both around sixteen billion. That is what makes it move so much.",
+ "source":"USDA WASDE, 12 August 2026. Carry-in is the figure implied by the published stocks and use.",
+ "steps":[{"label":"Carry-in","value":1945,"kind":"base"},
+          {"label":"Production","value":16013},
+          {"label":"Imports","value":25},
+          {"label":"Domestic use","value":-13055},
+          {"label":"Exports","value":-3275},
+          {"label":"Ending stocks","kind":"total"}]}
+```
+
+Stocks-to-use is the number a desk quotes out loud, because a bushel figure means nothing without the size of the market next to it. 1,653 over 16,330 is 10.1%.
+
+## Why the bottom line is a lever
+
+Look at the size of the terms. Production is 16.0 billion bushels. Ending stocks is 1.65 billion. Production is roughly **ten times** the line it feeds.
+
+So a 1% error in the crop is a 10% error in the carryout.
+
+Take the yield down two bushels, from 180.7 to 178.7. That is 1.1% on the yield line. On 88.6 million harvested acres it is 177 million bushels of production.
+
+| Yield (bu/ac) | Production | Ending stocks | Stocks-to-use |
+|---|---|---|---|
+| 178.7 | 15,836 | 1,476 | 9.0% |
+| 180.7 (USDA) | 16,013 | 1,653 | 10.1% |
+| 182.7 | 16,190 | 1,830 | 11.2% |
+
+```chart
+{"type":"bar","unit":"million bushels",
+ "title":"Two bushels of yield, eleven percent of carryout",
+ "x":["178.7 bu/ac","180.7 bu/ac","182.7 bu/ac"],
+ "series":[{"name":"Ending stocks","values":[1476,1653,1830]}],
+ "caption":"A 1.1% change in the yield input moves ending stocks by about 11% in either direction. The balance sheet is a lever, and the fulcrum sits very close to the crop.",
+ "source":"Calculated on the August 2026 WASDE sheet, holding demand and area constant at 88.6 m harvested acres."}
+```
+
+Eleven percent either way, out of a one percent input. That is the whole reason a yield estimate is worth arguing about for six weeks.
+
+**A warning about that table, though.** It holds demand fixed, and a real balance sheet does not. A sheet is a set of quantities *at an assumed price*. Cut the crop, lift the price, and exports and ethanol grind both start to fall — which rebuilds part of the stocks you just removed. Rationing is not a metaphor. It is the arithmetic by which the sheet closes at a higher price.
+
+## Stocks-to-use bends
+
+The relationship between stocks-to-use and price is not a line. It is a curve, and it bends.
+
+From 15% down to 12%, price barely notices. There is plenty of corn either way and the marginal bushel is not scarce. From 10% down to 8%, price goes vertical, because at that level somebody has to be rationed out of the market and price is the only tool that does it.
+
+Which means the same two-bushel yield move is worth almost nothing at a comfortable stocks-to-use and an enormous amount at a tight one. At 10.1%, corn is close enough to the bend that the market is paying attention — and that, rather than the headline figure, is why an August yield print gets traded as hard as it does.
+
+## The line that hides the sins
+
+**Feed and residual.** Two words, and the second is doing real work.
+
+Nobody counts the corn a hog eats. Feed use is inferred: take the quarterly Grain Stocks survey, subtract everything that can actually be measured, and whatever remains gets called feed and residual. So every measurement error in the sheet — in production, in exports, in the survey itself — lands in that one line.
+
+Which means that when feed and residual moves, sometimes it is telling you about livestock, and sometimes it is telling you the crop was never the size they said it was.
+
+Ending stocks is a residual. Feed and residual is a residual sitting inside it.
+
+## Where analysts actually disagree
+
+Here is what the argument sounds like the morning before a report.
+
+> **Analyst:** I have got carryout at one six eight.
+> **Trader:** On what yield?
+> **Analyst:** One eighty-one. Same as theirs.
+> **Trader:** Then where are we actually different?
+> **Analyst:** Exports. You are carrying three three, I am at three two.
+> **Trader:** A hundred million bushels of Mexico. That is not a crop view, that is a freight view.
+
+Notice where the disagreement sat. Not the yield — they agreed on the yield.
+
+That is the pattern, and it has a structural cause. **Supply is surveyed.** Fields are walked, ears are counted, acres are measured from satellite imagery and from farm programme filings. Two analysts working from the same surveys land close together.
+
+**Demand is inferred.** Nobody surveys a feedlot's intentions. Export commitments are known, but shipment timing is a decision somebody else has not made yet — which is precisely what the Chinese soybean programme is demonstrating this month.
+
+So two competent analysts converge on supply and diverge on demand. Which is why the interesting positions ahead of a report are rarely outright: a demand view is a view about *when* and *where*, and that lives in spreads.
+
+## Trading the surprise, not the number
+
+On 12 August USDA cut the corn yield to 180.7 bu/ac. The trade average going in was 182.5. So the yield surprise was 1.8 bushels — 159 million bushels of production, gone.
+
+And the carryout printed only about 72 million below what the trade was carrying.
+
+| Line | Effect |
+|---|---|
+| Yield surprise, −1.8 bu/ac × 88.6 m ac | −159 m bu |
+| Area and demand revisions | +87 m bu |
+| **Carryout surprise** | **−72 m bu** |
+
+Half the surprise disappeared on the way down the page. This is the trap that catches anyone who trades one line: **the market does not price the yield, it prices the carryout**, and the carryout has an area term, a demand term and a carry-in term, any of which can pay for a bad yield.
+
+December corn added roughly 20¼¢ on the print. It then did not hand it back — it closed Monday 17 August at 489½, above its report-day close. But that was a different trade. The follow-through came from impaired Black Sea export capacity, a soaked eastern belt and a soybean complex pulling the whole board higher on Chinese buying. None of it was in the report.
+
+Which is the last honest thing to say about report days. The number is the shock. The market is what happens over the next three weeks, and it needs its own reason.

package/ep07.script.txt

@@ -0,0 +1,117 @@
+Two analysts. Same report, same spreadsheet, opposite conclusions. ||| 0.4
+Neither of them has made an arithmetic mistake. ||| 0.6
+This is Soft Commodity Trading, episode seven. Today, the balance sheet. ||| 0.4
+The one-page object every grain desk argues about, and the report that updates it. ||| 0.8
+First, the tape. ||| 0.5
+Monday belonged to soybeans. November beans closed at twelve dollars sixteen, up twenty-three and a half cents. ||| 0.4
+That is a one and three quarter percent day, in a market that had already run all of last week. ||| 0.5
+December corn added six and a quarter, to four eighty-nine and a half. ||| 0.4
+Kansas City September wheat added four and a half cents to seven fifty-eight and three quarters. Chicago wheat took profits. ||| 0.6
+Two things did the buying. ||| 0.35
+China took another hundred and thirty-six thousand tonnes of U S beans. ||| 0.4
+And the crop split in two. The northwestern belt is dry. The eastern belt is too wet. ||| 0.5
+Monday afternoon's conditions had corn at sixty percent good to excellent, beans at sixty-one. ||| 0.4
+Both down a point on the week. Beans are seven points below where they were a year ago. ||| 0.6
+Now the political read, and today it is not the Black Sea. ||| 0.4
+It is the buyer. ||| 0.5
+China's purchases of new-crop U S beans have been running far ahead of anything that has actually sailed. ||| 0.4
+Vessel data through the end of July had U S to China departures averaging around eleven thousand tonnes a day. ||| 0.4
+That is close to nothing. The commitments are large. The loadings are not. Not yet. ||| 0.5
+That gap is an instrument, not an accident. ||| 0.4
+A commitment is a promise on a balance sheet. A loading is a fact on a vessel. ||| 0.4
+The buyer chooses when the promise becomes the fact, and that timing is worth more than the price they paid. ||| 0.7
+Which is convenient, because a commitment lands on exactly one line of the object we are building today. ||| 0.8
+So. The balance sheet. ||| 0.4
+Supply on top. Demand underneath. What is left over at the bottom. ||| 0.5
+Supply is three lines. Carry-in, production, imports. ||| 0.4
+Carry-in is what was still sitting there when the previous marketing year ended. ||| 0.4
+A marketing year for U S corn runs September to August. Beans the same. Wheat starts in June. ||| 0.4
+It is the crop's calendar, not the accountant's. ||| 0.5
+Production is the line everyone watches. Harvested acres times yield per harvested acre. ||| 0.4
+Note the word harvested. Not planted. ||| 0.4
+About eight million planted U S corn acres never get cut for grain in a normal year. ||| 0.5
+Imports, for U S corn, are a rounding error. Twenty-five million bushels. ||| 0.6
+Demand is where the money is, and it is four lines. ||| 0.4
+Feed and residual. Ethanol. Food, seed and other industrial. And exports. ||| 0.5
+Then ending stocks. Supply minus use. ||| 0.4
+It is not measured. It is what is left. ||| 0.7
+Let me build the real one, with the August numbers. ||| 0.5
+Eighty-eight point six million harvested acres. A hundred and eighty point seven bushels to the acre. ||| 0.4
+That is sixteen billion and thirteen million bushels of production. ||| 0.5
+Add imports of twenty-five million, and a carry-in of about one point nine four five billion. ||| 0.4
+Total supply, seventeen point nine eight three billion bushels. ||| 0.6
+Now take the demand out. ||| 0.35
+Domestic use, thirteen point zero five five billion. Exports, three point two seven five billion. ||| 0.4
+Total use, sixteen point three three billion. ||| 0.5
+Ending stocks. One point six five three billion bushels. ||| 0.7
+And now the number the desk actually quotes out loud. ||| 0.4
+One point six five three, over sixteen point three three. Ten point one percent. ||| 0.4
+Stocks to use. Ten point one. ||| 0.7
+Here is the thing about that bottom line. ||| 0.4
+It is the difference between two numbers that are both around sixteen billion. ||| 0.5
+Production is roughly ten times the size of the ending stocks it feeds. ||| 0.4
+So a one percent error in the crop is a ten percent error in the carryout. ||| 0.6
+Move the yield by two bushels. Just two, on a hundred and eighty. ||| 0.4
+That is a bit over one percent on the yield line. ||| 0.4
+It is a hundred and seventy-seven million bushels of production. ||| 0.4
+And it takes ending stocks to one point four seven six, or up to one point eight three. ||| 0.5
+Eleven percent either way, out of a one percent input. ||| 0.5
+Stocks to use swings from nine point zero to eleven point two. ||| 0.6
+That is the whole reason a yield estimate is worth arguing about for six weeks. ||| 0.4
+The balance sheet is a lever, and the fulcrum sits very close to the crop. ||| 0.8
+One more thing about stocks to use, and this is the part people quote wrongly. ||| 0.5
+The relationship to price is not a line. It is a curve, and it bends. ||| 0.5
+From fifteen percent down to twelve, price barely notices. There is plenty of corn either way. ||| 0.5
+From ten down to eight, price goes vertical. ||| 0.4
+Because at that level somebody has to be rationed out of the market, and price is the only tool that does it. ||| 0.5
+So the same two-bushel yield move is worth almost nothing at a comfortable stocks to use. ||| 0.4
+And it is worth an enormous amount at a tight one. ||| 0.5
+Ten point one is close enough to the bend that the market is paying attention. ||| 0.8
+Now the line that hides the sins. ||| 0.4
+Feed and residual. Two words, and the second one is doing real work. ||| 0.5
+Nobody counts the corn a hog eats. ||| 0.4
+Feed use is inferred. You take the quarterly stocks survey, subtract everything you can actually measure, and whatever remains gets called feed and residual. ||| 0.5
+So every measurement error in production, in exports, in the survey itself, lands in that one line. ||| 0.5
+Which means that when feed and residual moves, sometimes it is telling you about livestock. ||| 0.4
+And sometimes it is telling you the crop was never the size they said it was. ||| 0.7
+Ending stocks is a residual. We established that in episode two. ||| 0.4
+Feed and residual is a residual sitting inside the residual. ||| 0.6
+This is what the disagreement sounds like, the morning before a report. ||| 0.5
+ANALYST: I have got carryout at one six eight. ||| 0.25
+TRADER: On what yield? ||| 0.25
+ANALYST: One eighty-one. Same as theirs. ||| 0.25
+TRADER: Then where are we actually different? ||| 0.25
+ANALYST: Exports. You are carrying three three. I am at three two. ||| 0.25
+TRADER: A hundred million bushels of Mexico. That is not a crop view, that is a freight view. ||| 0.6
+Listen to where that argument actually sat. ||| 0.4
+Not the yield. They agreed on the yield. ||| 0.4
+They disagreed about demand. ||| 0.5
+And that is the pattern. Supply gets surveyed. Fields are walked, ears are counted, acres are measured from satellites and from farm program filings. ||| 0.5
+Demand is inferred. Nobody surveys a feedlot's intentions. ||| 0.5
+So two competent analysts converge on supply and diverge on demand. ||| 0.4
+Which is why the interesting positions before a report are rarely outright. ||| 0.4
+They are in the spreads, because that is where a demand view actually lives. ||| 0.8
+Last piece. Trading the surprise, not the number. ||| 0.5
+On the twelfth of August the U S D A cut the corn yield to a hundred and eighty point seven. ||| 0.4
+The trade average going in was a hundred and eighty-two and a half. ||| 0.4
+So the yield surprise was one point eight bushels, bearish for supply. A hundred and fifty-nine million bushels, gone. ||| 0.5
+And yet the carryout printed only about seventy-two million below what the trade was carrying. ||| 0.6
+Half the surprise disappeared on the way down the page. ||| 0.4
+Acres went up. Exports went up. The sheet absorbed it. ||| 0.6
+This is the trap that catches anybody who trades one line. ||| 0.4
+The market does not price the yield. It prices the carryout. ||| 0.4
+And the carryout has an area term, a demand term and a carry-in term, any of which can pay for a bad yield. ||| 0.7
+December corn jumped about twenty cents on the print. ||| 0.4
+And then it did not hand it back, because the Black Sea kept lifting the whole complex behind it. ||| 0.5
+It closed Monday at four eighty-nine and a half. Higher than it was on report day. ||| 0.6
+Which is the last honest thing to say about report days. ||| 0.4
+The number is the shock. The market is what happens over the next three weeks. ||| 0.8
+So. What to keep. ||| 0.5
+The balance sheet is not a forecast. It is an accounting identity, and it has to close. ||| 0.5
+Ending stocks is not measured. It is what is left, and that is exactly what makes it a lever on the crop. ||| 0.5
+About ten times the leverage for corn. Work that ratio out yourself for every commodity you trade, because it is different for every one. ||| 0.6
+Analysts agree on supply and fight about demand, because supply is counted and demand is inferred. ||| 0.5
+And the trade is never the print. It is the gap between the print and what was already in the price. ||| 0.7
+Tomorrow, the soybean complex and the crush. One seed, three markets, and the margin that runs the industry. ||| 0.5
+The quiz is in the notes. Today you build a small sheet and flex it, then go back to corn demand and to the physical chain. ||| 0.4
+Work them with a calculator, not from memory. ||| 0.5

package/feed.xml

@@ -18,6 +18,15 @@
       <title>Soft Commodity Trading</title>
       <link>https://storage.googleapis.com/podcast-audio-2647223968/index.html</link>
     </image>
+    <item>
+      <title>Ep 7 — WASDE and Building a Balance Sheet</title>
+      <description>How a grain balance sheet is built line by line, and why ending stocks — the line nobody measures — moves about ten times faster than the crop itself. Plus feed and residual, the line that hides the sins, and why two competent analysts agree on supply and fight about demand.</description>
+      <itunes:summary>How a grain balance sheet is built line by line, and why ending stocks — the line nobody measures — moves about ten times faster than the crop itself. Plus feed and residual, the line that hides the sins, and why two competent analysts agree on supply and fight about demand.</itunes:summary>
+      <enclosure url="https://cdn.jsdelivr.net/npm/@sdelsad/commodity-desk-daily@1.0.20/ep07.mp3" length="7381773" type="audio/mpeg"/>
+      <guid isPermaLink="false">https://cdn.jsdelivr.net/npm/@sdelsad/commodity-desk-daily@1.0.20/ep07.mp3</guid>
+      <pubDate>Tue, 18 Aug 2026 05:00:00 GMT</pubDate>
+      <itunes:duration>615</itunes:duration>
+    </item>
     <item>
       <title>Ep 6 — Corn, Crop Calendars and Weather Risk</title>
       <link>https://storage.googleapis.com/podcast-audio-2647223968/commodity-desk-daily/ep06.html</link>

package/glossary.md

@@ -9,6 +9,7 @@ Units, conventions and desk expressions, accumulated as the show introduces them
 - **asset-light** — renting elevators, terminals and plants rather than owning them  _(ep 2)_
 - **at** — the small word that introduces the offer side (462 bid, at 462 and a half)  _(ep 1)_
 - **bag (coffee)** — 60 kg, how the coffee trade counts volume  _(ep 1)_
+- **balance sheet** — the one-page supply and demand statement for one crop and one marketing year, built so that supply minus use equals ending stocks and the page closes  _(ep 7)_
 - **bid** — the price a buyer will pay  _(ep 1)_
 - **bill of lading** — receipt, contract of carriage and document of title in one, whoever holds it owns the cargo  _(ep 4)_
 - **bushel** — volume measure standardized into weight, 60 lb for soybeans and wheat, 56 lb for corn  _(ep 1)_
@@ -17,11 +18,13 @@ Units, conventions and desk expressions, accumulated as the show introduces them
 - **cancelling date** — the last day of the laycan, after which the counterparty may cancel  _(ep 4)_
 - **carry market (contango)** — a curve with later months above nearer ones, the market pays for storage  _(ep 3)_
 - **carry-in** — stocks left over from the previous season, the starting point of a balance sheet  _(ep 2)_
+- **carryout** — ending stocks, the desk's one-word name for what is left at the end of the marketing year  _(ep 7)_
 - **cents per bushel** — Chicago grain quoting unit, 4.39 dollars per bushel is spoken four thirty-nine  _(ep 1)_
 - **CFR** — cost and freight, the seller pays the voyage to a named destination but risk still passes at loading  _(ep 4)_
 - **charter party** — the contract hiring the vessel, between charterer and shipowner  _(ep 4)_
 - **CIF** — cost insurance and freight, CFR plus the seller buys the marine insurance the buyer would claim on  _(ep 4)_
 - **conversion factors** — 36.7 bushels per tonne for wheat and beans and 39.4 for corn, so cents per bushel times 0.367 or 0.394 gives dollars per tonne  _(ep 1)_
+- **Crop Production** — the USDA report published alongside WASDE carrying the survey-based yield and area figures  _(ep 7)_
 - **cross-hedge** — hedging with a contract that is not your grade or your origin, which removes flat price and adds correlation risk  _(ep 5)_
 - **cwt** — hundredweight, 100 lb, the quoting unit for US rice and cattle  _(ep 1)_
 - **cwt (hundredweight)** — 100 lb, the quoting unit for US rice  _(ep 1)_
@@ -37,6 +40,7 @@ Units, conventions and desk expressions, accumulated as the show introduces them
 - **durum** — the pasta wheat, a separate species with its own thin market  _(ep 5)_
 - **ethanol grind** — the rate at which ethanol plants consume corn, which slows when the plant margin turns negative and removes corn demand in steps  _(ep 6)_
 - **falling number** — the sprout-damage test, a low number demotes milling wheat to feed wheat  _(ep 5)_
+- **feed and residual** — the inferred demand line that carries livestock feeding together with every measurement error in the rest of the sheet  _(ep 7)_
 - **feed floor** — the price at which feed substitution demand appears under a grain, corn setting the floor under feed wheat  _(ep 6)_
 - **feed wheat** — wheat sold on energy and protein rather than milling specification, priced relationally against corn rather than at a flat price  _(ep 6)_
 - **firm** — a tradable quote that binds if accepted, often with a time limit  _(ep 1)_
@@ -46,11 +50,13 @@ Units, conventions and desk expressions, accumulated as the show introduces them
 - **FOB** — free on board, the cargo is priced at the load port with the buyer taking it from the ship's rail  _(ep 2)_
 - **front month** — the nearest actively traded contract month, where liquidity is deepest  _(ep 3)_
 - **full carry** — storage plus interest per month of holding grain, the practical ceiling on a carry spread  _(ep 3)_
+- **Grain Stocks** — the quarterly USDA survey of physical inventories, from which the feed and residual line is backed out  _(ep 7)_
 - **hard red spring (HRS)** — the 13.5 percent plus Minneapolis wheat bought to lift the protein of a grist  _(ep 5)_
 - **hard red winter (HRW)** — the 11 to 12.5 percent bread wheat priced at Kansas City, the US wheat that competes with the Black Sea  _(ep 5)_
 - **harvested acres** — area actually cut for grain, roughly 8 million acres below planted for US corn, and the denominator that yield is quoted against  _(ep 6)_
 - **hit** — your bid was taken by a seller  _(ep 1)_
 - **hit the bid** — to sell into someone else's bid  _(ep 1)_
+- **implied disappearance** — use derived by subtraction rather than by measurement, the technique that produces the residual lines of a balance sheet  _(ep 7)_
 - **inclusion rate** — the share of a single ingredient in a feed ration, capped by nutrition and by anti-nutritional factors  _(ep 6)_
 - **Incoterms** — the standard three-letter trade terms that allocate cost and risk between buyer and seller  _(ep 4)_
 - **indication** — a guide price that is not firm  _(ep 1)_
@@ -64,12 +70,16 @@ Units, conventions and desk expressions, accumulated as the show introduces them
 - **limit move** — an exchange-set maximum daily price change, trading pauses beyond it  _(ep 3)_
 - **line-up** — the queue of vessels waiting to load at a port, a key driver of origin basis  _(ep 2)_
 - **lot** — one futures contract, 5,000 bushels for Chicago grains, the unit desks count positions in  _(ep 1)_
+- **marketing year** — the accounting year a crop is measured in, September to August for US corn and soybeans and June to May for US wheat  _(ep 7)_
 - **Matif milling wheat (EBM)** — the Paris contract, 50 tonnes a lot quoted in euros per tonne and delivered into Rouen and Dunkirk  _(ep 5)_
 - **metric tonne** — 2,204.6 lb, the grain trading weight unit outside the US  _(ep 1)_
 - **month codes** — F G H J K M N Q U V X Z for January through December, the Z is December  _(ep 1)_
+- **NASS** — USDA's National Agricultural Statistics Service, the body running the surveys behind the published numbers  _(ep 7)_
+- **new crop** — the marketing year about to begin, priced by the contract months that follow the coming harvest  _(ep 7)_
 - **nomination** — formally naming the performing vessel under a cargo contract  _(ep 4)_
 - **notice of readiness (NOR)** — the master's formal declaration that the vessel has arrived and is ready, it starts the laytime clock  _(ep 4)_
 - **offer** — the price a seller will accept  _(ep 1)_
+- **old crop** — the marketing year now ending, priced by the contract months before the new harvest arrives  _(ep 7)_
 - **paper** — exchange futures and options, used by a physical desk to hedge rather than to speculate  _(ep 2)_
 - **physical (cash)** — real cargoes under contract with specs and load windows, as opposed to paper  _(ep 2)_
 - **planted acres** — area sown, the number that moves on farmer decisions and USDA area surveys  _(ep 6)_
@@ -92,6 +102,9 @@ Units, conventions and desk expressions, accumulated as the show introduces them
 - **test weight** — the density measure telling a miller how much flour comes out of a tonne  _(ep 5)_
 - **tick** — smallest price increment, a quarter cent per bushel in Chicago grains, worth 12.50 dollars per lot  _(ep 1)_
 - **ticker** — the short screen code a contract is spoken by, ZW wheat, ZC corn, ZS soybeans, ZM meal, ZL oil, KC coffee, SB sugar, CT cotton  _(ep 3)_
+- **total supply** — carry-in plus production plus imports, the top block of a balance sheet  _(ep 7)_
+- **total use** — domestic use plus exports, the bottom block of a balance sheet  _(ep 7)_
+- **trade average** — the published mean of analysts' pre-report estimates, and therefore the expectation already contained in the price  _(ep 7)_
 - **trend yield** — the yield a crop would produce on normal weather, the baseline against which a weather premium is measured  _(ep 6)_
 - **variation margin** — the daily cash settlement of a position mark to market, paid the same day  _(ep 3)_
 - **war-risk premium** — an insurance surcharge on a vessel's hull value for sailing into a conflict zone, quoted as a percentage  _(ep 2)_
@@ -100,6 +113,7 @@ Units, conventions and desk expressions, accumulated as the show introduces them
 - **washout** — cancelling two offsetting physical contracts by settling the price difference instead of shipping  _(ep 1)_
 - **weather premium** — the gap between where a crop trades and where it would trade at trend yield, the price of a distribution of outcomes rather than of a forecast  _(ep 6)_
 - **weather working day** — a laytime day that counts only when weather permits cargo work  _(ep 4)_
+- **whisper number** — the expectation the market is actually trading into a report, which can sit away from the published trade average  _(ep 7)_
 - **work** — leave an order resting with a broker  _(ep 1)_
 - **work an order** — leave an order resting at your price and wait  _(ep 1)_
 - **workable** — the quoted price is negotiable  _(ep 1)_

package/package.json

@@ -1,7 +1,7 @@
 {
   "name": "@sdelsad/commodity-desk-daily",
-  "version": "1.0.19",
-  "description": "Soft Commodity Trading - Ep 6: Corn, Crop Calendars and Weather Risk",
+  "version": "1.0.21",
+  "description": "Soft Commodity Trading - Ep 7: WASDE and Building a Balance Sheet",
   "license": "CC-BY-4.0",
   "keywords": [
     "podcast",

package/ep06.md

@@ -1,218 +0,0 @@
-# Market pulse
-
-**Wheat ran away with the week, and corn could not keep up.**
-
-| Commodity | Contract | Price | Change |
-|---|---|---|---|
-| Corn | Sep (CBOT) | 459 c/bu | +11¢ |
-| Corn | Dec (CBOT) | 477½ c/bu | +5½¢ |
-| Soybeans | Sep (CBOT) | 1177¾ c/bu | +11¾¢ |
-| Soybean meal | Sep (CBOT) | $310.20/short ton | +$2.80 |
-| Soybean oil | Sep (CBOT) | 69.44 c/lb | +65 pts |
-| Wheat SRW | Sep (CBOT) | 674¾ c/bu | +22¢ |
-| Wheat SRW | Dec (CBOT) | 679 c/bu | +10¾¢ |
-| Wheat HRW | Dec (KC) | 747¼ c/bu | +12½¢ |
-| Milling wheat | Spot (Matif) | €228.25/t | near July peak |
-
-Friday closed higher across the board, but the week belonged to wheat: more than four percent, the best week since mid-July. Corn added five and a half cents in December and no more.
-
-The reason is a supply story that finally found a buyer. Ukrainian drone strikes suspended operations at all three grain terminals at Novorossiysk, Russia's main Black Sea outlet. Russian August loadings are tracking near 2.5 million tonnes, under half the five-year seasonal pace and the weakest August since 2016/17. Ukraine's own month-to-date shipments are 201,700 t, down 76% year on year, with the deepwater corridor effectively shut since 22 July. Russia rejected a proposed partial ceasefire covering civilian shipping.
-
-The transmission is flow substitution, and it is one step further along than last week. Ten days ago both exporters were constrained and Chicago still finished unchanged, because no importer had actually moved its business. This week the buying started to appear elsewhere, and the price moved with it. Constraint alone does not lift a market. A buyer changing origin does.
-
-Corn's lag is not indifference. Corn demand is switchable in a way that milling wheat demand is not, and that is today's subject.
-
-```chart
-{"type":"line","mode":"index","unit":"index, Tue 11 Aug = 100",
- "title":"Wheat led, corn stalled after WASDE",
- "x":["Tue 11","Wed 12","Thu 13","Fri 14"],
- "series":[{"name":"Corn Dec (CBOT)","values":[460.50,480.75,477.75,477.50]},
-           {"name":"Wheat SRW Sep (CBOT)","values":[630.25,652.75,651.25,674.75]}],
- "caption":"Both jumped on the 12 August WASDE. Only wheat kept going, on Black Sea supply; corn gave the report back and sat still.",
- "source":"CBOT settlements, 11–14 August 2026, as reported in the daily market wraps."}
-```
-
-# Key takeaways
-
-- Corn is a demand story. Roughly two fifths of the US crop is feed and a large further slice is ethanol, and both of those buyers walk away at a price. Milling wheat demand does not.
-- Substitution runs in both directions. Corn is the floor under feed wheat, and feed wheat is the ceiling over corn.
-- A weather premium decays on the calendar, not on the forecast. Once pollination is over, no forecast changes the ear count.
-- Being right on the weather and wrong on the timing loses money — and being right on yield can still lose to acreage.
-- The safrinha is decided in October, in a soybean field, not in April in a corn field.
-
-# Vocabulary
-
-| Term | Meaning |
-|---|---|
-| **Safrinha** | Brazil's second corn crop, planted Feb–Mar into soybean stubble, about three quarters of Brazilian corn |
-| **Weather premium** | The gap between where a crop trades and where it would trade at trend yield, priced as a distribution rather than a forecast |
-| **Trend yield** | The yield a crop would produce on normal weather, the baseline a weather premium is measured against |
-| **Planted acres** | Area sown, the number that moves on farmer decisions and USDA area surveys |
-| **Harvested acres** | Area actually cut for grain, roughly 8 million acres below planted for US corn |
-| **Feed wheat** | Wheat sold on energy and protein rather than milling spec, priced against corn |
-| **Ration** | The formulated feed mix a mill grinds, in which each ingredient has an inclusion limit and a substitution price |
-| **Ethanol grind** | The rate at which ethanol plants consume corn, which slows when the plant margin turns negative |
-| **Distillers grains (DDGS)** | The protein co-product of ethanol production, sold back into feed |
-| **Abandonment** | Planted area never harvested for grain, from drought, flood or a switch to silage |
-
-# Quiz
-
-**Q1.** December corn settles at 477½ c/bu and December Chicago wheat at 679 c/bu. Convert both to dollars per tonne, then say how far wheat would have to fall, in cents per bushel, before a feed mill would begin swapping wheat for corn. Assume wheat earns a 4% feeding credit over corn.
-
-**Q2.** A trader is long December corn from mid-July on a drought forecast. The drought was real: USDA cuts the national yield from 183 to 180.7 bu/ac on 12 August. The trader is flat on the trade by 14 August. Explain, with numbers, how a correct yield call produced no money.
-
-**Q3.** It is late October. Rain has delayed soybean planting across Mato Grosso by roughly three weeks. Which corn contract does this argue for trading, in which direction, and what is the chain of causation? Name the month in which the damage, if it comes, actually happens.
-
-**Q4.** Ep 5 taught the cross-hedge. Today's mill is long 30,000 t of feed wheat bought flat to corn, and hedges it by selling Chicago corn futures rather than wheat futures. The Black Sea story then lifts Chicago wheat 30 c/bu while corn moves 5 c/bu. Where is the P&L, and was the hedge choice wrong?
-
-**Q5.** Ep 5 gave KC over Chicago at 68 c/bu, or about $26/t, as the protein spread. Today December KC is 747¼ and December Chicago 679, a spread of 68¼ c/bu. The spread has not moved while both markets rallied over 4%. What does an unchanged spread tell you about what kind of buying drove the week?
-
-**Q6.** Ep 3 taught that the curve pays for storage up to full carry. US corn harvest begins in September. If the December–March corn spread is trading at 60% of full carry today, what is the market telling a farmer with on-farm storage, and how would that message change if the Black Sea story pulled export demand forward into the fourth quarter?
-
-**Q7.** Ep 3 taught that a hedge converts price risk into liquidity risk. The mill in Q4 is short corn futures. Corn rallies 30 c/bu over two weeks on a late-season weather scare. Its physical wheat is worth more too. Why might the treasurer still be in trouble, and what is the number that matters?
-
-**Q8 — Conversion drill.** A Paraná safrinha forecast calls for an overnight low of 9 °C. On the same day a southern Plains station prints 113 °F. Convert each into the other scale, and say which of the two is the market-relevant number for a corn desk in August.
-
-# SOLUTIONS (spoilers)
-
-**A1.** Corn is 56 lb/bu, so a tonne is 39.368 bu; wheat is 60 lb/bu, so a tonne is 36.744 bu. Two different divisors — this is the trap.
-
-| | c/bu | bu/t | $/t |
-|---|---|---|---|
-| Corn Dec | 477.5 | 39.368 | 188.0 |
-| Wheat SRW Dec | 679.0 | 36.744 | 249.5 |
-
-With a 4% feeding credit, wheat's switching threshold is 188.0 × 1.04 ≈ $195/t. Wheat trades at $249.5/t, so the gap is $54.50/t. Converting back to the wheat bushel: 54.50 ÷ 36.744 = 148 c/bu. Chicago wheat would have to fall to roughly 531 c/bu before a feeder even looks. On a mill grinding 20,000 t/month at 60% corn — 12,000 t of corn — switching today would cost about $654,000 a month. Nobody switches. The point is the size of the distance: the feed floor is a real level, but it is a long way below a milling market in a supply scare.
-
-**A2.** The yield call was right and the production call was wrong, because yield is only one of the two terms.
-
-| Line | Value |
-|---|---|
-| July production | 16.000 bn bu |
-| Yield effect: −2.3 bu/ac × 88.6 m harvested ac | −204 m bu |
-| Area effect: acres found | +217 m bu |
-| August production | 16.013 bn bu |
-
-USDA lifted planted corn area to 96.73 m acres, part of about 2.8 m extra acres found across corn and beans. So a genuine 2.3 bu yield cut — the largest the trade expected — arrived alongside a production *increase*, and the second-largest crop on record. Dec corn added roughly 20 c on the print, gave most of it back over the following two sessions, and finished the week at 477½. The trap: a weather premium is a claim on the balance sheet, not on the yield line, and the balance sheet has an area term, a demand term and a carry-in term that can each pay for the yield.
-
-**A3.** Buy the contract that carries Brazilian second-crop supply — in Chicago terms the July contract, and on the Brazilian domestic curve the mid-year months. The instinct to trade the *near* corn on a Brazilian planting delay is the trap: nothing happens to supply for seven months.
-
-The causation runs: wet October delays soybean planting → beans harvest late, into late February → safrinha goes into the ground past its planting window → pollination slips from April into May → May is the start of the Brazilian dry season in the centre-west. So a three-week delay in October raises the probability of a pollination failure seven months later. The damage, if it comes, happens in **May**. The trade is to buy the corn contract that carries Brazilian second-crop supply — CBOT July, or the Brazilian domestic curve — while the market is still pricing a normal safrinha. The lesson underneath: the largest single risk to the world corn balance is set by a planting date in a different crop.
-
-**A4.** This is a cross-hedge and it behaved exactly as ep 5 warned. The mill bought physical feed wheat priced *flat to corn* — so its purchase price is contractually tied to corn — and hedged with corn futures. That part is internally consistent.
-
-| Leg | Move | P&L on 30,000 t |
-|---|---|---|
-| Physical wheat, priced flat to corn | +5 c/bu corn-equivalent | +$41,000 |
-| Short corn futures | +5 c/bu | −$41,000 |
-| Net | | ≈ 0 |
-
-(30,000 t of wheat ÷ 36.744 = 816,300 bu-equivalent; 5 c/bu ≈ $41,000.)
-
-The hedge was *not* wrong, because the exposure being hedged was corn — the contract said so. The mill did, however, forgo the 30 c/bu that outright wheat length would have made, which is not a hedging loss but the price of not having a view. Where it would go wrong is basis: if the mill's next purchase reprices against wheat rather than corn, the corn hedge stops matching the exposure and the 25 c/bu divergence between the two markets becomes real slippage — about $204,000 on this tonnage.
-
-**A5.** An unchanged KC–Chicago spread through a 4% rally says the buying was **not** quality-driven. A protein squeeze — a hard-wheat supply problem, a milling spec shortage — widens KC over Chicago, because it bids the protein and leaves the soft wheat behind. Here both classes rose together and the spread sat at roughly 68 c/bu.
-
-That is the signature of a *world balance sheet* move rather than a *quality* move: an export-capacity constraint at a competing origin lifts the whole US wheat complex, because the substitute for absent Black Sea tonnage is any exportable wheat, of any class. The desk reads it as flat-price risk, not spread risk, and hedges accordingly. If the spread had widened 20 c while both rallied, the story would have been protein, and the trade would have been in the spread, not the outright.
-
-**A6.** At 60% of full carry the market is paying most, but not all, of the cost of storage. Ep 3's rule: compare the spread to the full-carry ceiling and act on the gap.
-
-- **Farmer with on-farm storage**, whose marginal cost is interest and shrink rather than a commercial elevator tariff: 60% of full carry usually covers his real cost. Store, and sell the March against it — he is capturing a spread the market is paying him for.
-- **Farmer paying commercial storage** at the full tariff: 60% does not cover it. Sell at harvest, or sell the cash and buy a call if he wants the upside.
-
-If Black Sea disruption pulled export demand forward into Q4, the December contract would bid relative to March, the spread would compress toward zero or invert, and the message flips completely. An inverse tells the farmer the market wants the grain *now* and will not pay him to hold it. Holding through an inverse pays storage costs and loses the spread — punished twice, as ep 3 put it.
-
-**A7.** Because the hedge converted price risk into liquidity risk, and the two legs settle on different clocks. The short corn futures position takes variation margin in cash, daily, before the physical gain is realised in any invoice.
-
-Position size: 30,000 t ÷ 39.368 bu/t ≈ 762,000 bu of corn-equivalent, or about 152 contracts of 5,000 bu. A 30 c/bu adverse move on the short leg is 762,000 × $0.30 ≈ **$229,000** of variation margin, paid out in cash within days.
-
-The physical wheat is indeed worth more, but that value arrives when the ration is sold, or later. The number that matters is not the P&L — which is roughly flat — it is the peak cash drawdown against the committed credit line. This is the 2022 European wheat episode in miniature: solvent hedges, insolvent treasuries.
-
-**A8 — Conversion drill answer.** 9 °C → (9 × 9/5) + 32 = 16.2 + 32 = **48.2 °F**. 113 °F → (113 − 32) × 5/9 = 81 × 5/9 = **45 °C**.
-
-The market-relevant number in August is the 113 °F Plains print. The Paraná low is a Southern Hemisphere winter reading with no crop in the critical window — safrinha corn is already harvested by August and the new full-season crop is not yet planted. The Plains heat, by contrast, lands on late-filling row crops and on hard red winter wheat country's soil moisture going into autumn seeding. Same day, two temperatures, one of them tradeable.
-
-# Written edition
-
-## Corn is a demand story
-
-Corn is the largest crop on earth and almost none of it is eaten by people directly. Roughly two fifths of the US crop goes to feed. A large further slice goes to ethanol. Exports take about 3.275 billion bushels this marketing year. Food and industrial use is the smallest piece.
-
-That composition is the whole personality of the market. A miller who needs milling wheat needs milling wheat: the specification is the product, and there is no substitute at any price. A feeder needs neither corn nor wheat. A feeder needs energy and protein at the lowest cost per unit, subject to inclusion limits. To a nutritionist, corn and feed wheat are the same product. To a miller they are different products.
-
-Ethanol adds a second switch, and it is a margin rather than a preference. A plant buys corn and sells ethanol and distillers grains. When that margin goes negative the plant slows its grind, and a chunk of corn demand simply stops appearing at the bid. The demand curve for corn is not a smooth line. It has steps in it, and the steps are where the substitutes and the plant margins sit.
-
-## Pricing the switch
-
-Take Friday's settlements. The conversion is the first place people lose money, because corn and wheat do not share a divisor: corn is 56 lb to the bushel, so a tonne is 39.368 bushels; wheat is 60 lb, so a tonne is 36.744.
-
-| | c/bu | $/t |
-|---|---|---|
-| Corn, Dec | 477½ | 188.0 |
-| Wheat SRW, Dec | 679 | 249.5 |
-| Wheat HRW, Dec (KC) | 747¼ | 274.5 |
-| Feed-switch threshold for wheat | — | ~195 |
-
-Wheat carries a modest feeding credit against corn — more protein, comparable energy, offset by inclusion limits. Call it 4%. So wheat displaces corn at around $195/t. It is at $249.50. The gap is $54.50 a tonne.
-
-For a mill grinding 20,000 t a month at 60% corn inclusion, that is 12,000 t of corn, and switching to wheat would cost about $654,000 a month. The switch is not marginal. It is not close.
-
-Run it backwards and the same arithmetic gives the distance in bushel terms: $54.50/t ÷ 36.744 = 148 c/bu, so Chicago wheat would need to trade near 531 before a feeder picked up the phone. That number — the feed floor — is the reason wheat has a bottom in a bear market that has nothing to do with millers.
-
-```chart
-{"type":"bar","unit":"USD per tonne",
- "title":"Wheat is far above its feed value",
- "categories":["Corn Dec","Feed switch level","Wheat SRW Dec","Wheat HRW Dec"],
- "series":[{"name":"USD/t","values":[188.0,195.0,249.5,274.5]}],
- "caption":"A feeder switches from corn to wheat at about $195/t. Chicago wheat is $54 above that, so the substitution bid is nowhere near being triggered — wheat is trading as a milling market, not a feed market.",
- "source":"CBOT and KC settlements, 14 August 2026, converted at 39.368 bu/t corn and 36.744 bu/t wheat. Switch level assumes a 4% feeding credit."}
-```
-
-The relationship also runs the other way, which is the part people get backwards. If feed wheat is what corn competes with, then corn has a ceiling as well as a floor beneath wheat. Corn would have to reach roughly $240/t — about 609 c/bu — before today's wheat became the cheaper feed. At that point the substitution bid arrives from the other side and caps corn. **Corn is the floor under feed wheat; feed wheat is the ceiling over corn.**
-
-This is also why quotes in the feed market are relational rather than absolute. A feed buyer does not ask where wheat is. He asks where wheat is *against corn* — "eight over", "flat to corn" — because his spreadsheet only understands the difference. He has a ration, not a view.
-
-## The clock
-
-Corn has a weather market nearly every month of the year, but never the same corn.
-
-| Origin | Planting | Critical window | Harvest |
-|---|---|---|---|
-| United States | Apr–May | Pollination, mid-July | Sep–Nov |
-| Ukraine | Apr–May | July | Sep–Oct |
-| Brazil, full season | Sep–Oct | Dec–Jan | Feb–Mar |
-| Brazil, safrinha | Feb–Mar | Pollination, Apr–May | Jun–Aug |
-| Argentina | Sep–Dec | Dec–Feb | Mar–Jul |
-
-Two things in that table matter more than they look. First, the United States and Ukraine sit on the same calendar. Two origins in the same hemisphere with the same critical window are not a diversified supply base — one hemispheric heat pattern can hit both, which is precisely what makes a Northern Hemisphere July the single most consequential month in the corn year.
-
-Second, about three quarters of Brazilian corn is the safrinha, the second crop, drilled into soybean stubble as the beans come off. And the safrinha's dominant risk is not its own weather. It is the soybean harvest date sitting in front of it.
-
-The chain is mechanical. A wet October delays soybean planting. Late planting pushes the bean harvest into late February. A late bean harvest pushes safrinha drilling past its window. And a late safrinha pollinates in May, which is when the centre-west dry season starts. The bet on Brazilian corn is placed in October, in a soybean field, and settled in May.
-
-## The anatomy of a weather premium
-
-A weather premium is not a forecast. It is the price of a distribution: the gap between where corn trades and where it would trade at trend yield.
-
-It builds when the models disagree, and it builds fastest in the ten to fourteen days before a critical window, because that is when the market is buying optionality on an outcome that is still genuinely open. Then it does the thing that makes weather bulls miserable: **it decays whether or not the weather improves.** Time itself removes the possibility of damage. Once the crop has pollinated, no forecast changes the ear count. The calendar is short the weather bull every single day.
-
-Which is why being right on the weather and wrong on the timing loses money — and why, this month, being right on the weather and right on the yield also lost money.
-
-```chart
-{"type":"waterfall","unit":"million bushels",
- "title":"The drought was paid for out of acres",
- "caption":"A 2.3 bu/ac yield cut removed 204 million bushels. USDA then found enough planted area to more than replace it, and the crop got bigger.",
- "source":"USDA WASDE and Crop Production, 12 August 2026.",
- "steps":[{"label":"July production","value":16000,"kind":"base"},
-          {"label":"Yield −2.3 bu/ac","value":-204},
-          {"label":"Area found","value":217},
-          {"label":"August production","kind":"total"}]}
-```
-
-On 12 August USDA cut the national corn yield to 180.7 bu/ac from 183 — a real cut, the largest the trade was looking for, and the direct consequence of heat and dryness through pollination. On roughly 88.6 million harvested acres, 2.3 bushels is 204 million bushels gone.
-
-Production went up anyway. USDA raised planted area to 96.73 million acres, part of about 2.8 million extra acres found across corn and soybeans, and the crop printed 16.013 billion bushels — the second largest on record. December corn added around 20 cents on the day, gave most of it back over the next two sessions, and closed the week at 477½.
-
-Hidden inside that is a unit distinction worth owning. **Planted acres are not harvested acres.** Around 8 million acres of planted US corn never get cut for grain — silage, abandonment, flood, replant failure. Yield is quoted per *harvested* acre; area news moves on *planted*. Two different denominators in the same sentence, and a balance sheet that only balances if you keep them apart.
-
-The general lesson: a weather premium is a claim on the balance sheet, not on the yield line. Area, demand and carry-in can each pay for a bad yield, and this month area paid in full.

package/ep06.script.txt

@@ -1,119 +0,0 @@
-Corn is the biggest crop in the world, and almost nobody eats it. ||| 0.5
-Most of it is eaten by animals, or burned in an engine. ||| 0.5
-That one fact changes how the whole market trades. ||| 0.7
-This is Soft Commodity Trading, episode six. Corn, crop calendars, and weather risk. ||| 0.8
-First, the tape. ||| 0.5
-Friday closed higher across the board in Chicago. ||| 0.4
-September corn four fifty-nine, up eleven cents. ||| 0.4
-September beans eleven seventy-seven and three quarters, up eleven and three quarters. ||| 0.4
-September Chicago wheat six seventy-four and three quarters, up twenty-two cents. ||| 0.5
-Wheat did the work. Up more than four percent on the week. ||| 0.5
-December Kansas City finished at seven forty-seven and a quarter. ||| 0.6
-Here is why. Ukrainian drones hit Novorossiysk again, and this time all three grain terminals stopped. ||| 0.5
-Russian August loadings are tracking near two and a half million tonnes. ||| 0.4
-That is under half the five-year pace for August, and the weakest August since twenty sixteen. ||| 0.5
-Ukraine's own month-to-date shipments are down seventy-six percent, at two hundred thousand tonnes. ||| 0.5
-The deepwater corridor has been effectively shut since the twenty-second of July. ||| 0.5
-And Russia rejected a partial ceasefire that would have protected civilian shipping. ||| 0.6
-Watch the mechanism, not the headline. Both large sellers are constrained at the same time. ||| 0.5
-Last week the market shrugged, because no buyer had actually switched origin. ||| 0.5
-This week they started to. That is the difference between a threat and a flow. ||| 0.7
-Corn came along for the ride, December at four seventy-seven and a half. ||| 0.5
-Which is a good place to ask a question. ||| 0.4
-Why is corn so slow to follow wheat up? ||| 0.7
-Because corn is a demand story, and demand has a price at which it says no. ||| 0.6
-Start with where corn goes. ||| 0.4
-Roughly two fifths of the American crop is feed. Another large slice is ethanol. ||| 0.4
-Exports take about three and a quarter billion bushels this year. ||| 0.4
-Food and industrial use is the smallest piece. ||| 0.5
-Feed and fuel are both switchable. That is the whole point. ||| 0.5
-Ethanol has its own switch, and it is a margin, not a taste. ||| 0.4
-A plant buys corn and sells ethanol and distillers grains. ||| 0.4
-When that margin goes negative it slows the grind, and corn demand simply stops showing up. ||| 0.5
-So the demand curve for corn is not smooth. It has steps in it. ||| 0.6
-A wheat buyer who needs milling wheat needs milling wheat. ||| 0.4
-A feeder does not need corn. A feeder needs energy and protein. ||| 0.6
-So corn and feed wheat are the same product to a nutritionist, and different products to a miller. ||| 0.7
-Let's price that. ||| 0.5
-Corn is fifty-six pounds to the bushel, so a tonne is about thirty-nine point four bushels. ||| 0.4
-Wheat is sixty pounds, so a tonne is about thirty-six point seven. ||| 0.4
-Two different divisors. This is where people lose the trade. ||| 0.6
-December corn at four seventy-seven and a half is one hundred eighty-eight dollars a tonne. ||| 0.5
-December Chicago wheat at six seventy-nine is two hundred forty-nine and a half a tonne. ||| 0.6
-Now, wheat carries a small feeding credit against corn. More protein, similar energy. ||| 0.4
-Call it four percent. So wheat has to be at or below about one ninety-five to displace corn. ||| 0.6
-It is at two forty-nine. The gap is fifty-four dollars a tonne. ||| 0.7
-Take a mill grinding twenty thousand tonnes a month, sixty percent corn. ||| 0.4
-That is twelve thousand tonnes of corn a month. ||| 0.4
-Switching to wheat would cost that mill six hundred and fifty-four thousand dollars a month. ||| 0.5
-Nobody switches. ||| 0.6
-So how far would wheat have to fall? ||| 0.4
-Fifty-four dollars a tonne is a hundred and forty-eight cents a bushel. ||| 0.4
-Chicago wheat would need to trade near five thirty. ||| 0.5
-That is the distance between the milling market and the feed floor. ||| 0.7
-Here is how it sounds when the gap actually closes. ||| 0.5
-BROKER: Feed wheat, September, delivered mill. I can work you a number. ||| 0.25
-FEEDER: Where against corn? ||| 0.25
-BROKER: Call it eight over on a tonne basis. ||| 0.25
-FEEDER: Eight over is not a switch. Bring me flat and I'll take three cargoes. ||| 0.25
-BROKER: Flat to corn, you take thirty thousand? ||| 0.25
-FEEDER: Flat to corn, I take thirty and I re-formulate this week. ||| 0.7
-Notice what he did not say. He never named a wheat price. ||| 0.5
-He named a relationship. Eight over corn, or flat to corn. ||| 0.5
-The feeder does not have a view on wheat. He has a ration and a spreadsheet. ||| 0.7
-Now turn it around, because this is the part people get backwards. ||| 0.5
-If wheat is the thing corn competes with, then corn has a ceiling too. ||| 0.5
-Corn would have to reach about two hundred forty dollars a tonne to make today's wheat competitive. ||| 0.4
-That is six dollars nine cents a bushel. ||| 0.6
-So corn is the floor under feed wheat, and feed wheat is the ceiling over corn. ||| 0.5
-Two markets, one substitution, working in both directions. ||| 0.7
-Now the clock. ||| 0.5
-Corn has a weather market almost every month of the year, but never the same corn. ||| 0.6
-America plants in April and May, pollinates in the middle of July, harvests from September. ||| 0.5
-Ukraine does the same thing on the same calendar. ||| 0.4
-That matters more than it sounds. The same heat dome can hit both. ||| 0.4
-Two origins on one hemisphere is not diversification. ||| 0.6
-Then the south. Brazil's full-season crop plants in September and comes off in February. ||| 0.4
-But three quarters of Brazil's corn is the safrinha. The second crop. ||| 0.5
-It goes into the ground in February and March, straight into soybean stubble as the beans come off. ||| 0.4
-It pollinates in April and May, and it harvests through the middle of the year. ||| 0.6
-Argentina plants from September and harvests March to July. ||| 0.5
-So the desk's quiet weeks are November through January, and even then it is watching South American planting. ||| 0.7
-Here is the part that is not obvious. ||| 0.5
-The safrinha's biggest risk is not its own weather. ||| 0.5
-It is the soybean harvest date in front of it. ||| 0.6
-A wet October delays soybean planting. A late planting delays the bean harvest into February. ||| 0.4
-A late bean harvest pushes safrinha planting past its window. ||| 0.4
-And a late safrinha pollinates in May, which is the start of the Brazilian dry season. ||| 0.6
-So the real bet on Brazilian corn is made in October, in a soybean field. ||| 0.7
-Which brings us to the weather premium itself. ||| 0.5
-A weather premium is not a forecast. It is the price of a distribution. ||| 0.5
-It is the gap between where corn trades and where corn would trade at trend yield. ||| 0.6
-It builds when the models disagree, and it builds fastest ten to fourteen days before the critical window. ||| 0.5
-Then it does something people find painful. ||| 0.5
-It decays whether or not the weather improves. ||| 0.6
-Because time itself removes the possibility of damage. ||| 0.4
-Once the corn has pollinated, no forecast on earth changes the ear count. ||| 0.5
-The calendar is short the weather bull, every single day. ||| 0.7
-Look at what just happened. ||| 0.5
-U S D A cut the corn yield to one eighty point seven, from one eighty-three. ||| 0.4
-That is a real cut. Two point three bushels off the national yield. ||| 0.5
-On roughly eighty-eight and a half million harvested acres, that is two hundred million bushels gone. ||| 0.6
-And production went up. ||| 0.6
-Because U S D A found acres. Planted area went to ninety-six point seven million. ||| 0.5
-Two point eight million extra acres across corn and beans. ||| 0.4
-The crop came in at sixteen point zero one three billion bushels. Second largest on record. ||| 0.7
-That is the weather premium being paid for out of area. ||| 0.5
-A trader who was right on the drought, and right on the yield cut, still made almost nothing. ||| 0.6
-And here is the unit that hid it. Planted acres are not harvested acres. ||| 0.4
-Around eight million acres of planted corn never get cut for grain. ||| 0.4
-Silage, abandonment, flood. Yield is quoted per harvested acre, area moves on planted. ||| 0.5
-Two different denominators in the same sentence. ||| 0.7
-So what do you actually remember. ||| 0.6
-Corn is a demand story. Feed and fuel both walk away at a price. ||| 0.5
-Substitution runs both ways. Corn floors feed wheat, feed wheat caps corn. ||| 0.5
-The weather premium decays on the calendar, not on the forecast. ||| 0.5
-And the safrinha is decided by the soybean harvest before it. ||| 0.7
-Tomorrow, the report that stops every grain desk. W A S D E, and how to build a balance sheet you can actually flex. ||| 0.5
-The quiz is in the notes, with a conversion drill and full solutions. ||| 0.5
-This has been Soft Commodity Trading. ||| 0.5