Track record

Every pick we've made.
Including the bad ones.

This page is append-only. Nothing is ever deleted, edited or quietly hidden after the fact — which means the losing weeks are here too, because a record without them isn't a record.

  • 4,180 settled predictions
  • Updated nightly
  • Losing months included
67.3%cut rate, all time (2,813 of 4,180)
+8.4%return on investment at flat stakes
4,180settled predictions on record
2losing months out of 11
±1.9points — worst calibration error

How to read this honestly

A 67% cut rate sounds impressive and tells you almost nothing on its own — a model that only ever backed heavy favourites would beat it and still lose money. The number that matters is the second one: return at flat stakes. That's what you'd have actually made staking the same amount on every pick, and it's the only figure that survives contact with a bookmaker's margin.

The third number matters too, in the other direction. Two of our eleven months were losing months. June was down 4.1%. That isn't a bug or a bad patch we're apologising for — it's what a genuine 8% long-run edge looks like month to month. Anyone showing you eleven green months is showing you a filtered sample.

Recent settled SharpCut predictions with odds, stated probability, confidence and result
DateFixtureMarket OddsOur prob.Conf.Result
Sat 02 AugArsenal v ChelseaPremier League · 3-1Over 2.51.8571%HIGHWon
Sat 02 AugDortmund v MainzBundesliga · 2-0Dortmund win1.3579%HIGHWon
Sat 02 AugReal Betis v SevillaLa Liga · 1-1Under 2.52.0553%MEDWon
Sat 02 AugCeltic v RangersScottish Prem · 1-1Over 2.51.6863%MEDLost
Sat 02 AugPSV v FeyenoordEredivisie · 3-2Over 2.51.5566%HIGHWon
Fri 01 AugNapoli v RomaSerie A · 1-0BTTS1.7260%MEDLost
Fri 01 AugInter v LecceSerie A · 3-0Inter −1 AH1.6266%HIGHWon
Thu 31 JulAjax v TwenteEredivisie · 4-1Over 2.51.4870%HIGHWon
Thu 31 JulLyon v NiceLigue 1 · 1-0Under 2.51.9555%MEDWon
Wed 30 JulMan City v SevillaChampions League · 1-1Man City win1.4269%MEDLost
Wed 30 JulBayern v LazioChampions League · 3-1Over 2.51.6067%HIGHWon
Tue 29 JulBoston v MiamiNBA · 118-104Over 214.51.9056%MEDWon
Tue 29 JulDenver v PhoenixNBA · Den +2Denver −4.51.8755%MEDLost
Mon 28 JulAlcaraz v RuneATP Toronto · 6-3 6-4Alcaraz 2-02.1051%MEDWon
Mon 28 JulBrighton v WolvesPremier League · 2-2BTTS1.6664%HIGHWon
Sun 27 JulAtlético v OsasunaLa Liga · 2-0Atlético win1.4874%HIGHWon
Sun 27 JulLeverkusen v WerderBundesliga · 2-1Over 3.52.3046%LOWLost
Sun 27 JulJuventus v TorinoSerie A · 1-0Under 2.51.7262%MEDWon

Showing the last 18 settled predictions. In the product this table runs to every prediction ever published, filterable by sport, league, market, month and confidence band. At a flat ₦500 stake this sample returned +₦2,155 (ROI +23.9%).

Calibration is the real test

When we say 70%, does it happen 70% of the time? Every published prediction, bucketed by the probability we stated, against what actually occurred. The closer the line sits to the diagonal, the more honest the number on the card was.

Our worst bucket is out by 1.9 points. We slightly under-call the 80%+ band, which is the safer direction to be wrong in — but it's still an error and it's still published.

30%60%90% 30%60%90%
observed perfect n = 4,180

Now go and check the maths yourself.

The demo runs the same engine on sample fixtures — every probability computed in front of you.