Prediction Accuracy
After a round finishes, we score the predictions against actual fantasy points (FIFA feed) — error, ranking, scouting-bonus calibration, and how the recommended XI did. The hints feed back into tuning the model.
attacking-output multiplier the tuner learned per position (1.0 = no change)
predicted chance to clear threshold vs what actually happened
| Bucket | Predicted | Observed | n |
|---|---|---|---|
| 0-20% | 11% | 3% | 230 |
| 20-40% | 29% | 19% | 146 |
| 40-60% | 47% | 40% | 58 |
| 60-80% | 66% | 70% | 10 |
- XI actual points (capt ×2)
- 92.0
- Captain
- Kylian Mbappé (14 pts)
- Captain rank in XI
- #1 of 11
- Best captain (hindsight)
- Kylian Mbappé
- Avg player (field)
- 2.93 pts
where the model flatters players — chasing differentials is only right if the low-owned buckets aren't inflated
- →Over-predicting by 0.74 pts on average — consider scaling attacking rates / base goals down.
- →MID: over-predicting by 1.26 — adjust MID rates.
- →FWD: over-predicting by 1.29 — adjust FWD rates.
Every squad's real score vs the hindsight-best XIs, plus every big miss classified to the model factor that caused it — so each round makes the next projection better.
- DEFPedro Porro4.2 →16
DEF returned 16 vs 4.2 projected — clean-sheet/save/conceded points are systematically under-modelled for DEFs.
- MIDYouri Tielemans7.2 →18
Returned 18 vs 7.2 projected — team/fixture strength for this matchup was likely underrated.
- DEFMarc Cucurella4.1 →13
DEF returned 13 vs 4.1 projected — clean-sheet/save/conceded points are systematically under-modelled for DEFs.
- DEFKai Trewin0.3 →9
Model gave him only 5% to start, but he played and returned 9. The appearance model missed his role.
- DEFSidny Cabral0.4 →9
DEF returned 9 vs 0.4 projected — clean-sheet/save/conceded points are systematically under-modelled for DEFs.
- MIDDeroy Duarte2.3 →10
Low goal prior (1%) but hauled 10 — likely a finishing spike. Check role support before chasing.
- MIDOusmane Dembélé14.8 →6
Played but blanked (6 vs 14.8 projected) — mostly variance.
- FWDDeniz Undav9.8 →2
Ceiling pick projected 9.8 but returned 2. High-variance blank; the miss cost is amplified if captained.
- MIDRubén Vargas10.8 →3
Played but blanked (3 vs 10.8 projected) — mostly variance.
- MIDNathan Saliba8.7 →1
Ceiling pick projected 8.7 but returned 1. High-variance blank; the miss cost is amplified if captained.
- MIDJude Bellingham8.6 →1
Ceiling pick projected 8.6 but returned 1. High-variance blank; the miss cost is amplified if captained.
- MIDDésiré Doué8.4 →1
Ceiling pick projected 8.4 but returned 1. High-variance blank; the miss cost is amplified if captained.
- 8×Ceiling pick blanked — captaincy/exposure sizing, not projection
- 4×Defensive scoring calibration (clean sheets, saves, conceded)
- 2×Start-probability model (appearances / rotation signal)
- 2×One-off finishing spike — do NOT over-correct, check role support
- 2×Variance blank — played, no returns; only act if it repeats
- 1×Team strength / fixture odds for this matchup
- 1×Per-90 tackle / chance-created rates (volume routes)
backtested 4 Aug, 08:15 UTC