MatchLab XI

Read the pitch. Pick your XI.

Your match-day edge

Aryan's workspace

Single-user app, no sign-in. The demo match is synthetic and kept separate.

Gemini Flash · CricketData · Cricsheet · Open-Meteo

MatchLab XIAryan's workspace

Checking stored predictions and actual points…

Loading fixtures…

Checking stored predictions and actual points…

Learning Lab

A better process, every match.

Find repeated errors in stored predictions vs actual points, then approve or reject bounded weight changes.

Illustrative synthetic history · not measured performance

Reviewed matches

Synthetic demo history

Player results with actuals

MAE 7.77 pts

Bias (predicted − actual)

+2.62 pts

Positive bias = overprediction

Predicted vs actual player points (mean per match)

9 matches · 93 player results · synthetic

All-rounders overpredicted by 9.44 pts on average in this synthetic history.

All-rounders
n=24 · 8 matches · pred 49.9 vs actual 40.46 · Overpredicted by 9.4 pts on average; same direction in 8 of 8 matches
+9.44 pts
Captain / vice-captain picks
n=16 · 8 matches · pred 56.48 vs actual 52.13 · |bias| 4.4 < 6.8 pts threshold
+4.36 pts
Bowlers
n=27 · 9 matches · pred 41.94 vs actual 40.19 · direction consistent in 5/9 matches
+1.76 pts
Wicketkeepers
n=8 · 8 matches · pred 32.33 vs actual 33.13 · direction consistent in 4/8 matches
-0.8 pts
Batters
n=34 · 9 matches · pred 44.08 vs actual 44.79 · direction consistent in 5/9 matches
-0.72 pts

Minimum-evidence gateMet in demo (synthetic)

93/30 player results · 9/3 matches.

Synthetic records qualify only inside the demo. They never count toward real accuracy.

Bounded change to factor weights / role multipliers. Applies to future matches only.

No pending proposal

Evidence is sufficient and a pattern is flagged. Ask Gemini to summarise it and propose a bounded update. You decide.