fplkit 1.1.0__tar.gz → 1.2.0__tar.gz

{fplkit-1.1.0 → fplkit-1.2.0}/.agents/TOOLS.md

@@ -27,7 +27,7 @@ Errors use `emit_json_error()` with `{"command", "error"}` shape.
 | `init` | Set up fpl-cli with your FPL IDs and optional features | No | General | No | direct-api |
 | `status` | Show FPL gameweek status and upcoming deadlines | Yes | General | No | direct-api |
 | `fixtures` | Show fixtures for a gameweek | Yes | General | No | direct-api |
-| `player` | Look up a player's stats, xG, ownership and fixture run. JSON includes `ep_next` (predicted pts next GW) and `ep_this` (current GW); panel shows `xPts` (ep_next) | Yes | General | No | direct-api |
+| `player` | Look up a player's stats, xG, ownership and fixture run. JSON includes `ep_next` (predicted pts next GW) and `ep_this` (current GW); panel shows `xPts` (ep_next). `--detail` (`-d`) shows GW-by-GW performance and, for FWD/MID with sufficient history, xGI sustainability (per-match GI-xGI divergence -> form modifier). With `custom_analysis`: JSON adds `info.adjusted_npxg_per_90` (fixture-adjusted) and `info.raw_npxg_per_90` (Understat season avg); panel shows `adj. npxG/90: X.XXX (raw: Y.YYY)` | Yes | General | No | direct-api |
 | `stats` | List players with filtering and sorting. `--value` adds quality/value per £m columns; `--window N` sets rolling lookback (3-10) for `rolling_pts_per_m`. Sortable by `ep_next`/`ep_this` (FPL predicted points); both in JSON output | Yes | General | No | direct-api |
 | `history` | Show historical player performance across seasons | Yes | General | No | direct-api |
 | `league` | Show live league standings for Classic and Draft leagues | No | General | No | direct-api |
@@ -38,10 +38,10 @@ Errors use `emit_json_error()` with `{"command", "error"}` shape.
 | `preview` | Run full pre-gameweek analysis and generate report | No | General | No | via-agent |
 | `review` | Review a completed gameweek - squad performance and standings | No | General | No | via-agent |
 | `league-recap` | Recap a completed gameweek - awards, standings, and banter | No | General | No | via-agent |
-| `captain` | Analyse and rank captain options for next gameweek | Yes | Classic | Yes | via-agent |
-| `differentials` | Find differential picks - high potential, low ownership | Yes | Classic | Yes | via-agent |
-| `targets` | Find transfer targets - high performers across all ownership | Yes | Classic | Yes | via-agent |
-| `transfer-eval` | Compare transfer OUT player against IN candidates | Yes | General | Yes | via-agent |
+| `captain` | Analyse and rank captain options for next gameweek. JSON candidates include `adjusted_npxg_per_90` and `raw_npxg_per_90` when fixture adjustment is active. Consistency tiebreaker (CV-xGI percentile) phased in GW6-10 | Yes | Classic | Yes | via-agent |
+| `differentials` | Find differential picks - high potential, low ownership. Inverted consistency bonus (volatile players score higher, phased in GW6-10) | Yes | Classic | Yes | via-agent |
+| `targets` | Find transfer targets - high performers across all ownership. Consistency bonus (CV-xGI percentile, phased in GW6-10) | Yes | Classic | Yes | via-agent |
+| `transfer-eval` | Compare transfer OUT player against IN candidates. JSON includes `adjusted_npxg_per_90` and `raw_npxg_per_90` per player when fixture adjustment is active | Yes | General | Yes | via-agent |
 | `allocate` | Select mathematically optimal 15-player squad within budget | Yes | Classic | Yes | direct-api |
 | `waivers` | Show waiver recommendations for your draft league | Yes | Draft | Yes | via-agent |
 | `squad` | Analyse your FPL squad health and fixtures | Yes | General | No | via-agent |

{fplkit-1.1.0 → fplkit-1.2.0}/.gitignore

@@ -69,6 +69,7 @@ docs/plans/
 docs/brainstorms/
 docs/solutions/
 docs/ideas/
+docs/discarded/
 docs/positioning.md
 docs/releasing.md
 

fplkit-1.2.0/CHANGELOG.md

@@ -0,0 +1,33 @@
+# Changelog
+
+All notable changes to this project will be documented in this file.
+## [1.1.0] - 2026-04-07
+
+### Bug Fixes
+
+- wrap migration in try/except, update architecture.md paths
+- CLI hygiene batch (P10-P13) (#4)
+- Understat name matching for abbreviated/punctuated web_names
+- PR #4 code review follow-ups (#5)
+- review fixes - gather resilience, CSV field protection, test gaps
+- correct README data source count and Core-Insights repo URL
+
+### Features
+
+- add cached dataset fetcher with ETag conditional requests (#6)
+- add rolling pts/£m metric and rename value fields (#7)
+- expose ep_next/ep_this (FPL predicted points) (#8)
+- implement CoreInsightsClient for 2025-26 season data
+- add HistoricalDataProvider composition layer, update call sites
+
+### Refactoring
+
+- extract shared types, rename season helpers, separate cache dirs
+
+## [1.0.0] - 2026-04-03
+
+### Features
+
+- CI pipeline and PyPI publish for v1.0.0 (#1)
+
+

{fplkit-1.1.0 → fplkit-1.2.0}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: fplkit
-Version: 1.1.0
+Version: 1.2.0
 Summary: CLI tool for Fantasy Premier League analysis - classic and draft formats
 Project-URL: Homepage, https://github.com/rossgroomio/fpl-cli
 Project-URL: Repository, https://github.com/rossgroomio/fpl-cli

fplkit-1.2.0/config/fixture_predictions.yaml

@@ -0,0 +1,41 @@
+metadata:
+  last_updated: '2026-04-07'
+  notes: Updated from Ben Crellin BGW/DGW schedule (GW33-GW38)
+predicted_blanks:
+- gameweek: 34
+  teams:
+  - BHA
+  - BOU
+  - BUR
+  - CHE
+  - LEE
+  - MCI
+  confidence: confirmed
+- gameweek: 37
+  teams:
+  - BOU
+  confidence: low
+predicted_doubles:
+- gameweek: 33
+  teams:
+  - BHA
+  - BOU
+  - BUR
+  - CHE
+  - LEE
+  - MCI
+  confidence: confirmed
+- gameweek: 36
+  teams:
+  - CRY
+  - MCI
+  confidence: high
+- gameweek: 36
+  teams:
+  - BOU
+  confidence: low
+- gameweek: 37
+  teams:
+  - CRY
+  - MCI
+  confidence: low

{fplkit-1.1.0 → fplkit-1.2.0}/docs/architecture.md

@@ -361,8 +361,8 @@ Services live in `fpl_cli/services/` and provide the computation layer between a
 
 | Service | Purpose |
 |---|---|
-| `player_scoring` | Central scoring engine: `prepare_scoring_data()`, all score functions, `shrink_scores()` |
-| `player_prior` | Bayesian early-season confidence (GW1-10 shrinkage) |
+| `player_scoring` | Central scoring engine: `prepare_scoring_data()`, all score functions, `shrink_scores()`. Form modifiers: `compute_form_trajectory()` (direction over recent GWs) and `compute_xgi_sustainability()` (ATK-only rolling xGI divergence -> [0.85, 1.15] multiplier). Fixture-adjusted npxG: `compute_adjusted_npxg()` / `build_adjusted_npxg_lookup()` normalise historical xG by opponent Elo; `apply_adjusted_npxg()` overwrites `npxG_per_90` in agent enrichment when data is available. Consistency signals: `build_consistency_lookup()` computes 5 per-player signals (CV-xGI percentile, blank rate, floor percentile, involvement rate, GK consistency) with GW6-10 phase-in; additive bonuses per scoring family (target/waiver cv*1.5, differential inverted cv*0.75, lineup cv*0.75, bench floor*1.5+inv*0.75) |
+| `player_prior` | Bayesian early-season confidence (GW1-10 shrinkage); threads `PlayerProfile.reliability` (historical availability rate) through `PlayerPrior` to agents |
 | `team_ratings` | TeamRatingsService + Calculator (1-7 scale, 4 axes) |
 | `matchup` | Fixture matchup scoring (0-10), 3-GW recency-weighted |
 | `fixture_predictions` | BGW/DGW predictions from YAML + live detection |
@@ -416,7 +416,7 @@ All providers share the `LLMResponse` contract. `OpenAICompatProvider` supports
 | `FPLDraftClient` | FPL Draft API | Draft leagues, waivers, squad data |
 | `UnderstatClient` | understat.com | npxG, xA, xGChain, xGBuildup per-90 stats |
 | `VaastavClient` | vaastav/FPL GitHub | Historical CSV data (3 seasons: 2022-25), price trends, GW-level profiles |
-| `CoreInsightsClient` | Core-Insights/FPL GitHub | Current-season CSV data (2025-26+), season aggregates, GW trends |
+| `CoreInsightsClient` | Core-Insights/FPL GitHub | Current-season CSV data (2025-26+), season aggregates, GW trends, per-GW match-level xG + Elo (from `By Tournament/Premier League/GW{n}/`) |
 | `HistoricalDataProvider` | Composition layer | Unifies vaastav + Core-Insights via `make_historical_provider()` |
 | `FootballDataClient` | football-data.org | League standings, match results |
 | `FPLPriceScraper` | FPL website | Price change scraping (needs credentials) |
@@ -516,7 +516,7 @@ fpl_cli/
 │   ├── fpl.py                    # FPLClient (main API, caches bootstrap-static)
 │   ├── fpl_draft.py              # FPLDraftClient
 │   ├── understat.py              # UnderstatClient + match_fpl_to_understat()
-│   ├── historical_types.py       # Shared dataclasses (SeasonHistory, PlayerProfile, GwTrendProfile) + compute_trend/compute_acceleration
+│   ├── historical_types.py       # Shared dataclasses (SeasonHistory, PlayerProfile, GwTrendProfile) + compute_trend/compute_acceleration/compute_reliability
 │   ├── vaastav.py                # VaastavClient (historical seasons 2022-25 via DatasetFetcher)
 │   ├── core_insights.py          # CoreInsightsClient (current season 2025-26+ via DatasetFetcher)
 │   ├── historical.py             # HistoricalDataProvider (composition: vaastav + Core-Insights)
@@ -527,7 +527,7 @@ fpl_cli/
 │       ├── openai_compat.py      # OpenAICompatProvider (OpenAI, Groq, Together, Ollama)
 │       └── perplexity.py         # PerplexityProvider (extends OpenAICompat)
 ├── services/
-│   ├── player_scoring.py         # Scoring engines + prepare_scoring_data() + shrink_scores()
+│   ├── player_scoring.py         # Scoring engines + prepare_scoring_data() + shrink_scores() + consistency signals
 │   ├── player_prior.py           # Player prior (Bayesian early-season confidence)
 │   ├── team_ratings.py           # TeamRatingsService + Calculator (1-7 scale)
 │   ├── team_ratings_prior.py     # Pre-GW5 prior ratings for blending

{fplkit-1.1.0 → fplkit-1.2.0}/docs/command-reference.md

@@ -57,7 +57,7 @@ fpl captain --format json
 
 Output columns: Score, Atk, Def, Form±, Pos±.
 
-Score combines position-weighted matchup quality, recent form (with trajectory adjustment), and xGI per 90. DGW players are scored across both fixtures. See [Captain Score](custom-analysis.md#captain-score) for the full formula, [Matchup Scoring](custom-analysis.md#matchup-scoring) for column definitions.
+Score combines position-weighted matchup quality, recent form (with trajectory adjustment), xGI per 90, and a consistency tiebreaker (CV-xGI percentile, phased in GW6-10). DGW players are scored across both fixtures. See [Captain Score](custom-analysis.md#captain-score) for the full formula, [Matchup Scoring](custom-analysis.md#matchup-scoring) for column definitions.
 
 ### Transfer Targets
 
@@ -75,7 +75,7 @@ Groups players into tiers:
 - **Popular** (15-30% owned): Emerging picks
 - **Differential** (<15% owned): Low-ownership value
 
-Target score combines xG metrics, form, PPG, and 3-GW matchup quality, normalised to 0-100. Subject to [early-season shrinkage](custom-analysis.md#early-season-confidence-gw1-10). See [Target Score](custom-analysis.md#target-score) for the full formula.
+Target score combines xG metrics, form, PPG, 3-GW matchup quality, and consistency (CV-xGI percentile bonus, phased in GW6-10), normalised to 0-100. Subject to [early-season shrinkage](custom-analysis.md#early-season-confidence-gw1-10). See [Target Score](custom-analysis.md#target-score) for the full formula.
 
 ### Transfer Evaluation
 
@@ -98,6 +98,7 @@ Output columns:
 - **Fixtures** - next 3 opponents with FDR
 - **Form** - FPL form (last 30 days PPG)
 - **Status** - availability indicator
+- **Avail** - historical availability rate (recency-weighted starts across previous seasons). Null when no historical data.
 - **Quality** - price-independent player quality (0-100). Uses `VALUE_QUALITY_WEIGHTS`. Null when no Understat match.
 - **Value** - quality per GBP million (`quality_score / price`). Higher = more output per pound. Null when no Understat match or price is 0. *(classic only)*
 - **Price** - current price *(classic only)*
@@ -120,7 +121,7 @@ fpl differentials -m 200     # Require 200+ minutes played
 fpl differentials --format json  # JSON envelope (metadata: {gameweek})
 ```
 
-Differential score combines xG metrics, form, ownership bonus, and 3-GW matchup quality, normalised to 0-100. Subject to [early-season shrinkage](custom-analysis.md#early-season-confidence-gw1-10). See [Differential Score](custom-analysis.md#differential-score) for the full formula.
+Differential score combines xG metrics, form, ownership bonus, 3-GW matchup quality, and consistency (inverted CV-xGI bonus - volatile players score higher, phased in GW6-10), normalised to 0-100. Subject to [early-season shrinkage](custom-analysis.md#early-season-confidence-gw1-10). See [Differential Score](custom-analysis.md#differential-score) for the full formula.
 
 ### Waiver Recommendations
 
@@ -133,7 +134,7 @@ fpl waivers --format json
 
 Identifies squad weaknesses by position, ranks available free agents by waiver score, suggests who to drop for each pickup. This covers the waiver wire (unclaimed players) only - trade recommendations between managers are not in scope.
 
-Waiver score combines xGI, form, PPG, and 3-GW matchup quality, normalised to 0-100. Uses a stricter minutes factor than target/differential because draft waivers are a season commitment. Subject to [early-season shrinkage](custom-analysis.md#early-season-confidence-gw1-10). See [Waiver Score](custom-analysis.md#waiver-score) for the full formula.
+Waiver score combines xGI, form, PPG, 3-GW matchup quality, and consistency (CV-xGI percentile bonus, phased in GW6-10), normalised to 0-100. Uses a stricter minutes factor than target/differential because draft waivers are a season commitment. Subject to [early-season shrinkage](custom-analysis.md#early-season-confidence-gw1-10). See [Waiver Score](custom-analysis.md#waiver-score) for the full formula.
 
 ## Fixture & Strategic Planning
 
@@ -357,6 +358,7 @@ When a player has an Understat match, `fpl player` computes and displays two add
 | **pts_per_m** | `total_points / price` (per £m). Raw season points efficiency. |
 | **form_per_m** | `form / price` (per £m). Recent form efficiency. |
 | **rolling_pts_per_m** | Points per £m over the last N qualifying fixtures (configurable via `--window`). Captures recent form-adjusted value. |
+| **adj. npxG/90** | Fixture-adjusted non-penalty xG per 90: npxG normalised by opponent Elo over a rolling window. Shown alongside raw when the adjustment changes the value. See [Fixture-Adjusted npxG](custom-analysis.md#fixture-adjusted-npxg). |
 
 `quality_score` and `quality_per_m` are `null` when no Understat match exists. In JSON output (`--format json`), they appear under `info.quality_score` and `info.quality_per_m`. In the Rich panel, they appear as `Quality: 85 | Value: 11.3/£m`.
 
@@ -373,7 +375,7 @@ Both fields default to `0.0` for unavailable/injured players. In JSON output, th
 
 #### Player Detail Flags
 
-**`--detail` (`-d`)**: GW-by-GW match performance from the FPL API. Shows gameweek, opponent, minutes, goals, xG, assists, xA, and points for the last 10 matches.
+**`--detail` (`-d`)**: GW-by-GW match performance from the FPL API. Shows gameweek, opponent, minutes, goals, xG, assists, xA, and points for the last 10 matches. For FWD/MID players with sufficient recent history, also shows xGI sustainability: the per-match GI-xGI divergence and the resulting form modifier (e.g. `xGI Sustainability: +0.18/match -> 0.94x form`).
 
 **`--understat` (`-u`)**: Combined Understat analysis: shot analysis (total shots, shots on target, average xG per shot, body part split, situation breakdown) and situation profile. Includes a data-through date caveat since Understat data can lag behind the live season.
 

{fplkit-1.1.0 → fplkit-1.2.0}/docs/custom-analysis.md

@@ -301,11 +301,42 @@ Current-season data is blended with a prior from the previous season's Understat
 
 Stored in `config/team_ratings.yaml`.
 
+## Fixture-Adjusted npxG
+
+The scoring pipeline normally consumes npxG/90 as a flat season average from Understat. A player with a soft early schedule accumulates a high rate and also earns a strong matchup score — double-counting fixture difficulty.
+
+Fixture-adjusted npxG normalises historical xG by the Elo strength of each opponent before computing the per-90 rate. This isolates fixture difficulty to the matchup component alone.
+
+### How it works
+
+For each completed match within a 7-match / 12-GW rolling window:
+
+1. **npxG approximation**: `xg - (penalties_scored + penalties_missed) × 0.76` (removes penalty xG from the Core-Insights match total)
+2. **Adjustment factor**: `median_elo / opponent_elo`, capped at [0.80, 1.25] — a match against a 1900-Elo side is scaled up; a match against an 800-Elo side is scaled down
+3. **Adjusted per-90**: `(npxg × factor) / minutes × 90` per match, averaged over qualifying matches
+4. Players with fewer than 4 qualifying matches fall back to raw Understat npxG/90
+
+The median Elo is computed fresh from the season's match data each run. Elo ratings come from per-gameweek `matches.csv` files in the Core-Insights dataset (`By Tournament/Premier League/GW{n}/`).
+
+### Display
+
+When `custom_analysis: true` and match data is available, `fpl player` shows:
+
+```
+adj. npxG/90: 0.312 (raw: 0.385)
+```
+
+The existing `npxG` panel line continues to show the raw Understat season total. In JSON output, `info.adjusted_npxg_per_90` and `info.raw_npxg_per_90` are added when custom analysis is enabled. Captain and transfer evaluation agent outputs include both fields when an adjustment is active.
+
+If Core-Insights data is unavailable, the pipeline falls back to raw npxG/90 with no error surfaced to users.
+
 ## Quality & Value Scores
 
 Available via `fpl stats --value` and `fpl player` when Understat data exists.
 
-**quality_score** (0-100): Normalised player output quality using `VALUE_QUALITY_WEIGHTS`. Weights form and PPG heavily to capture current FPL points production rate. GK/DEF use a defensive variant (dc_per_90 replaces attacking xG stats).
+**quality_score** (0-100): Normalised player output quality using `VALUE_QUALITY_WEIGHTS`. Weights form and PPG heavily to capture current FPL points production rate. Position-specific scoring paths diverge for defensive players:
+- **GK**: dedicated signals via `for_gk()` weights — saves per 90, defensive quality (inverted xGC/90, range 0-2), and clean sheet rate. Normalised against `GK_VALUE_CEILING` (28.2).
+- **DEF**: `dc_per_90` (defensive contribution rate) replaces attacking xG stats via `without_xgi()`. Normalised against `VALUE_CEILING` (24.3).
 
 **quality_per_m**: `quality_score / price` (per £m). Within-position budget efficiency - higher means more output per pound. Not meaningful for cross-position comparison. Null when price is 0.
 

{fplkit-1.1.0 → fplkit-1.2.0}/fpl_cli/_version.py

@@ -18,7 +18,7 @@ version_tuple: tuple[int | str, ...]
 commit_id: str | None
 __commit_id__: str | None
 
-__version__ = version = '1.1.0'
-__version_tuple__ = version_tuple = (1, 1, 0)
+__version__ = version = '1.2.0'
+__version_tuple__ = version_tuple = (1, 2, 0)
 
 __commit_id__ = commit_id = None

{fplkit-1.1.0 → fplkit-1.2.0}/fpl_cli/agents/action/waiver.py

@@ -14,11 +14,15 @@ from fpl_cli.api.fpl import FPLClient
 from fpl_cli.api.fpl_draft import FPLDraftClient
 from fpl_cli.models.types import EnrichedPlayer, WaiverTarget
 from fpl_cli.services.player_scoring import (
+    ConsistencySignals,
+    apply_adjusted_npxg,
+    apply_consistency,
     apply_shrinkage,
     build_player_evaluation,
     calculate_waiver_score,
     compute_aggregate_matchup,
     compute_form_trajectory,
+    compute_xgi_sustainability,
     prepare_scoring_data,
 )
 
@@ -40,6 +44,8 @@ class WaiverAgent(Agent):
         super().__init__(config)
         self.client = FPLDraftClient()
         self.fpl_client = FPLClient()
+        self._adjusted_npxg_lookup: dict[int, float] | None = None
+        self._consistency_lookup: dict[int, ConsistencySignals] | None = None
         self.league_id = config.get("draft_league_id") if config else None
         self.entry_id = config.get("draft_entry_id") if config else None
 
@@ -170,10 +176,13 @@ class WaiverAgent(Agent):
             data = await prepare_scoring_data(
                 self.fpl_client, include_players=True,
                 include_history=True, include_prior=True,
+                include_match_data=True,
             )
             next_gw_id = data.next_gw_id
             self._player_histories = data.player_histories or {}
             self._player_priors = data.player_priors
+            self._adjusted_npxg_lookup = data.adjusted_npxg_lookup
+            self._consistency_lookup = data.consistency_lookup
             scoring_ctx = data.scoring_ctx
 
             # Enrich available players with matchup and FDR
@@ -322,11 +331,21 @@ class WaiverAgent(Agent):
         history = histories.get(player.get("id", 0), [])
         if history:
             enrichment["form_trajectory"] = compute_form_trajectory(history, next_gw_id)
+            sustainability, divergence = compute_xgi_sustainability(history, next_gw_id, player.get("position", "MID"))
+            enrichment["xgi_sustainability"] = sustainability
+            enrichment["xgi_divergence"] = divergence
         priors = getattr(self, "_player_priors", None)
         if priors:
             prior = priors.get(player.get("id", 0))
             if prior:
                 enrichment["prior_confidence"] = prior.confidence
+        if "npxG_per_90" not in enrichment:
+            npxg = player.get("npxG_per_90")
+            if npxg is not None:
+                enrichment["npxG_per_90"] = npxg
+        pid = int(player.get("id") or 0)
+        apply_adjusted_npxg(enrichment, pid, self._adjusted_npxg_lookup)
+        apply_consistency(enrichment, pid, self._consistency_lookup)
         evaluation, _ = build_player_evaluation(
             player,
             enrichment=enrichment,
@@ -497,6 +516,14 @@ class WaiverAgent(Agent):
                     target, drop_candidate, team_counts
                 )
 
+                priors = getattr(self, "_player_priors", None)
+                target_prior = priors.get(target.get("id", 0)) if priors else None
+                reliability = target_prior.reliability if target_prior is not None else None
+
+                con_lookup = getattr(self, "_consistency_lookup", None)
+                con_signals = con_lookup.get(target.get("id", 0)) if con_lookup else None
+                con_pct = con_signals.cv_xgi_percentile if con_signals is not None else None
+
                 rec: dict[str, Any] = {
                     "priority": len(recommendations) + 1,
                     "target": {
@@ -505,6 +532,8 @@ class WaiverAgent(Agent):
                         "position": pos,
                         "form": target.get("form"),
                         "waiver_score": target.get("waiver_score"),
+                        "reliability": reliability,
+                        "cv_xgi_percentile": con_pct,
                     },
                     "drop": {
                         "name": drop_candidate.get("player_name") if drop_candidate else None,

{fplkit-1.1.0 → fplkit-1.2.0}/fpl_cli/agents/analysis/bench_order.py

@@ -8,12 +8,16 @@ from fpl_cli.agents.base import Agent, AgentResult, AgentStatus
 from fpl_cli.api.fpl import FPLClient
 from fpl_cli.models.player import FORMATION_LIMITS, PlayerStatus
 from fpl_cli.services.player_scoring import (
+    ConsistencySignals,
     ScoringContext,
+    apply_adjusted_npxg,
+    apply_consistency,
     apply_shrinkage,
     build_fixture_matchups,
     build_player_evaluation,
     calculate_bench_score,
     compute_form_trajectory,
+    compute_xgi_sustainability,
     prepare_scoring_data,
 )
 
@@ -67,6 +71,7 @@ class BenchOrderAgent(Agent):
                 self.client,
                 include_players=True, include_understat=True,
                 include_history=True, include_prior=True,
+                include_match_data=True,
             )
             all_players = data.players or []
             player_map = {p.id: p for p in all_players}
@@ -74,6 +79,8 @@ class BenchOrderAgent(Agent):
             scoring_context = data.scoring_ctx
             team_map = data.team_map
             understat_by_id = data.understat_lookup
+            adjusted_npxg_lookup = data.adjusted_npxg_lookup
+            consistency_lookup = data.consistency_lookup
 
             bench_ids = context["bench"]
             bench_players = [player_map[pid] for pid in bench_ids if pid in player_map]
@@ -94,6 +101,8 @@ class BenchOrderAgent(Agent):
                     understat_by_id=understat_by_id,
                     player_histories=player_histories,
                     player_priors=data.player_priors,
+                    adjusted_npxg_lookup=adjusted_npxg_lookup,
+                    consistency_lookup=consistency_lookup,
                 )
                 scored_bench.append(score_data)
 
@@ -239,6 +248,8 @@ class BenchOrderAgent(Agent):
         understat_by_id: dict[int, dict[str, float]] | None = None,
         player_histories: dict[int, list[dict[str, Any]]] | None = None,
         player_priors: dict[int, Any] | None = None,
+        adjusted_npxg_lookup: dict[int, float] | None = None,
+        consistency_lookup: dict[int, ConsistencySignals] | None = None,
     ) -> dict[str, Any]:
         """Score a bench player via the scoring engine."""
         team = context.team_map.get(player.team_id)
@@ -251,11 +262,16 @@ class BenchOrderAgent(Agent):
             us_data = understat_by_id.get(player.id)
             if us_data:
                 enrichment.update(us_data)
+        apply_adjusted_npxg(enrichment, player.id, adjusted_npxg_lookup)
+        apply_consistency(enrichment, player.id, consistency_lookup)
 
         if player_histories:
             history = player_histories.get(player.id, [])
             if history:
                 enrichment["form_trajectory"] = compute_form_trajectory(history, next_gw_id)
+                sustainability, divergence = compute_xgi_sustainability(history, next_gw_id, player.position_name)
+                enrichment["xgi_sustainability"] = sustainability
+                enrichment["xgi_divergence"] = divergence
 
         if player_priors:
             prior = player_priors.get(player.id)

{fplkit-1.1.0 → fplkit-1.2.0}/fpl_cli/agents/analysis/captain.py

@@ -12,12 +12,16 @@ from fpl_cli.api.fpl import FPLClient
 from fpl_cli.models.player import Player
 from fpl_cli.models.types import CaptainCandidate
 from fpl_cli.services.player_scoring import (
+    ConsistencySignals,
     ScoringContext,
+    apply_adjusted_npxg,
+    apply_consistency,
     apply_shrinkage,
     build_fixture_matchups,
     build_player_evaluation,
     calculate_captain_score,
     compute_form_trajectory,
+    compute_xgi_sustainability,
     prepare_scoring_data,
 )
 
@@ -43,6 +47,8 @@ class CaptainAgent(Agent):
     def __init__(self, config: dict[str, Any] | None = None):
         super().__init__(config)
         self.client = FPLClient()
+        self._adjusted_npxg_lookup: dict[int, float] | None = None
+        self._consistency_lookup: dict[int, ConsistencySignals] | None = None
 
         # Differential thresholds
         self.differential_threshold = config.get("differential_threshold", 10.0) if config else 10.0
@@ -63,11 +69,14 @@ class CaptainAgent(Agent):
             self.client,
             include_players=True, include_understat=True,
             include_history=True, include_prior=True,
+            include_match_data=True,
         )
         all_players = data.players or []
         player_map = {p.id: p for p in all_players}
         self._player_histories = data.player_histories or {}
         self._player_priors = data.player_priors
+        self._adjusted_npxg_lookup = data.adjusted_npxg_lookup
+        self._consistency_lookup = data.consistency_lookup
 
         return (
             player_map,
@@ -216,6 +225,8 @@ class CaptainAgent(Agent):
             us_data = understat_by_id.get(player.id)
             if us_data:
                 enrichment.update(us_data)
+        apply_adjusted_npxg(enrichment, player.id, self._adjusted_npxg_lookup)
+        apply_consistency(enrichment, player.id, self._consistency_lookup)
 
         # xGI_per_90 fallback for players without Understat data
         minutes_safe = max(player.minutes, 1)
@@ -226,6 +237,9 @@ class CaptainAgent(Agent):
         history = histories.get(player.id, [])
         if history:
             enrichment["form_trajectory"] = compute_form_trajectory(history, next_gw_id)
+            sustainability, divergence = compute_xgi_sustainability(history, next_gw_id, player.position_name)
+            enrichment["xgi_sustainability"] = sustainability
+            enrichment["xgi_divergence"] = divergence
 
         # Bayesian prior confidence
         priors = getattr(self, "_player_priors", None)

{fplkit-1.1.0 → fplkit-1.2.0}/fpl_cli/agents/analysis/starting_xi.py

@@ -7,13 +7,17 @@ from typing import TYPE_CHECKING, Any
 from fpl_cli.agents.base import Agent, AgentResult, AgentStatus
 from fpl_cli.api.fpl import FPLClient
 from fpl_cli.services.player_scoring import (
+    ConsistencySignals,
     ScoringContext,
+    apply_adjusted_npxg,
+    apply_consistency,
     apply_shrinkage,
     build_fixture_matchups,
     build_player_evaluation,
     calculate_lineup_score,
     compute_aggregate_matchup,
     compute_form_trajectory,
+    compute_xgi_sustainability,
     prepare_scoring_data,
     select_starting_xi,
 )
@@ -67,12 +71,15 @@ class StartingXIAgent(Agent):
                 self.client,
                 include_players=True, include_understat=True,
                 include_history=True, include_prior=True,
+                include_match_data=True,
             )
             all_players = data.players or []
             player_map = {p.id: p for p in all_players}
             player_histories = data.player_histories or {}
             scoring_context = data.scoring_ctx
             understat_by_id = data.understat_lookup
+            adjusted_npxg_lookup = data.adjusted_npxg_lookup
+            consistency_lookup = data.consistency_lookup
 
             squad_ids = context["squad"]
             squad_players = [player_map[pid] for pid in squad_ids if pid in player_map]
@@ -95,6 +102,8 @@ class StartingXIAgent(Agent):
                     understat_by_id=understat_by_id,
                     player_histories=player_histories,
                     player_priors=data.player_priors,
+                    adjusted_npxg_lookup=adjusted_npxg_lookup,
+                    consistency_lookup=consistency_lookup,
                 )
                 scored.append(score_data)
 
@@ -150,6 +159,8 @@ class StartingXIAgent(Agent):
         understat_by_id: dict[int, dict[str, float]] | None = None,
         player_histories: dict[int, list[dict[str, Any]]] | None = None,
         player_priors: dict[int, Any] | None = None,
+        adjusted_npxg_lookup: dict[int, float] | None = None,
+        consistency_lookup: dict[int, ConsistencySignals] | None = None,
     ) -> dict[str, Any]:
         """Score a squad player via the scoring engine."""
         team = context.team_map.get(player.team_id)
@@ -167,11 +178,16 @@ class StartingXIAgent(Agent):
             us_data = understat_by_id.get(player.id)
             if us_data:
                 enrichment.update(us_data)
+        apply_adjusted_npxg(enrichment, player.id, adjusted_npxg_lookup)
+        apply_consistency(enrichment, player.id, consistency_lookup)
 
         if player_histories:
             history = player_histories.get(player.id, [])
             if history:
                 enrichment["form_trajectory"] = compute_form_trajectory(history, next_gw_id)
+                sustainability, divergence = compute_xgi_sustainability(history, next_gw_id, player.position_name)
+                enrichment["xgi_sustainability"] = sustainability
+                enrichment["xgi_divergence"] = divergence
 
         if player_priors:
             prior = player_priors.get(player.id)