ValueFinder Football | Gameweek 35 Preview
ValueFinder Beta — identifying mispriced probabilities, not predicting winners
What is ValueFinder Football?
ValueFinder Football uses expected goals (xG) data to model the true probability of each Premier League fixture going Over or Under 2.5 goals — and compares that against what the bookmaker odds are implying. Every team carries a set of data-driven labels updated each gameweek: traits like <High Scorer>, <Leaky Defence>, <Goals Drought> and <In Form Attack> that reflect current form rather than reputation.
Where the gap between our model probability and the market falls in our validated 4-6pp sweet spot, we flag it as a value selection. This gameweek there are no confirmed selections — but the fixture analysis below contains some of the most interesting model divergences we have seen so far. The model is working. The market is telling a different story on several fixtures. We let the data speak.
We are not predicting winners. We are identifying mispriced probabilities. Losses are variance. Closing Line Value and long-run calibration are the true indicators of edge.
Gameweek 35 Fixture Analysis
Leeds vs Burnley Friday 1 May
Leeds <High Scorer> · <Goals Streak> · <In Form Attack>
Burnley <Leaky Defence> · <Cold Attack>
H2H <H2H High Scoring>
Model O2.5: 72.8% | Market O2.5: 58.8% | Gap: +13.95% | BTTS: 50.0%
One of the larger gaps of the gameweek but sitting well above our 6pp cap. Leeds arrive carrying <High Scorer>, <Goals Streak> and <In Form Attack> — three corroborating labels pointing toward a high-scoring side in excellent current form.
Burnley’s <Leaky Defence> and <Cold Attack> combination suggests a side conceding heavily while struggling to create at the other end.
The <H2H High Scoring> history reinforces the picture. The model sees a genuinely goal-heavy fixture but the gap at +13.95pp means the market is likely pricing in information we can’t see — team news, tactical setup, motivational factors at both ends of the table. A fixture to watch rather than act on, but the underlying data is compelling.
Brentford vs West Ham Saturday 2 May
Brentford <High Scorer> · <High Variance> · <In Form Attack>
West Ham <In Form Attack>
Model O2.5: 47.7% | Market O2.5: 59.5% | Gap: -11.83% | BTTS: 52.5%
A significant divergence in the Under direction. The model sees only 47.7% Over probability while the market implies 59.5% — an 11.83pp gap suggesting the model thinks this game is considerably tighter than the bookmakers do.
Brentford carry <High Scorer> and <In Form Attack> which contradicts the low probability at first glance, but their <High Variance> label means their games swing between extremes.
West Ham bring <In Form Attack> but no attacking quality label to suggest they will unlock a Brentford defence. The model is seeing a closer, lower-scoring game than the market expects. Outside our validated band but worth noting if Under markets are of interest.
Newcastle vs Brighton Saturday 2 May
Newcastle <Leaky Defence> · <Goals Drought> · <In Form Attack> Brighton <High Scorer> · <High Variance> · <Cold Attack>
H2H <H2H Low Scoring>
Model O2.5: 65.8% | Market O2.5: 56.2% | Gap: +9.58% | BTTS: 46.9%
An interesting tension between the labels here. Newcastle carry <Leaky Defence> which would normally corroborate an Over, but their <Goals Drought> tells you their recent matches have consistently gone Under 2.5 despite the defensive vulnerability.
Brighton are <High Scorer> on rolling xG but <Cold Attack> in recent weeks — their output has dipped below their season average.
The <H2H Low Scoring> history between these two sides is the most telling detail — these teams have historically kept games tight regardless of form. The gap at +9.58pp sits above the cap. The model sees more goals than the market but the H2H and Brighton’s cold form make this less straightforward than the raw numbers suggest.
Wolves vs Sunderland Saturday 2 May
Wolves <Leaky Defence> · <High Variance> · <Cold Attack>
Sunderland <Cold Attack>
Model O2.5: 61.4% | Market O2.5: 46.7% | Gap: +14.65% | BTTS: 57.9%
The largest positive gap of the gameweek. Wolves carry <Leaky Defence> and <High Variance> — a combination that suggests unpredictable, potentially high-scoring games — while Sunderland bring only <Cold Attack> as a label, reflecting their limited attacking output in recent matches after a brilliant return season to the league.
The model sees meaningful Over probability at 61.4% but the market is pricing Under at 53.3% implied. Gap of +14.65pp is well above our cap. The discrepancy likely reflects Sunderland’s status as a newly promoted side — the market may be applying a promoted-side discount that needs to be removed after their debut season.
Bournemouth vs Crystal Palace Sunday 3 May
Bournemouth <High Scorer> · <In Form Attack>
Crystal Palace <High Variance> · <In Form Attack>
H2H <H2H Low Scoring>
⚑ Context: Crystal Palace — Conference League Thursday, rotation risk (-1)
Model O2.5: 72.2% | Market O2.5: 58.1% | Gap: +14.09% | BTTS: 71.5%
The context overlay is doing important work here. Crystal Palace are in European action on Thursday and rotation risk has been applied — a -1 conviction adjustment that would suppress any flag conviction even if the gap moved into the band.
Bournemouth carry <High Scorer> and <In Form Attack> and the model sees a high probability of goals at 72.2%. However the <H2H Low Scoring> history between these two sides is a genuine counterweight — and a rotated Palace side may actually produce a tighter, more defensive game than the xG ratings suggest. The 71.5% BTTS is the model’s most optimistic output of the gameweek. Gap above the cap, no flag, but a fixture where the European schedule context matters significantly.
Aston Villa vs Tottenham Sunday 3 May
Aston Villa <High Scorer> · <Goals Drought> · <In Form Attack>
Tottenham <Defensive Block> · <High Variance> · <Cold Attack>
H2H ⚔️ <H2H High Scoring>
⚑ Context: Aston Villa — Europa League Thursday, rotation risk (-1)
Model O2.5: 52.5% | Market O2.5: 56.8% | Gap: -4.31% | BTTS: 54.4%
The context overlay applies again here — Aston Villa in European action Thursday, rotation risk applied. Villa carry a contradictory label set: <High Scorer> on rolling xG but <Goals Drought> in recent matches, meaning their games have consistently gone Under despite strong underlying numbers. Tottenham’s <Defensive Block> and <Cold Attack> combination points to a side that is neither creating nor conceding much. The <H2H High Scoring> history is the one factor supporting Over but with rotation risk on Villa and Spurs in poor form, the model’s 52.5% feels more realistic than the market’s 56.8%. The -4.31pp gap is outside our band in the Under direction.
Chelsea vs Nottingham Forest Monday 4 May
Chelsea <Defensive Block> · <Cold Attack>
Nottingham Forest <Defensive Block> · <Cold Attack>
⚑ Context: Nottingham Forest — Europa League Thursday, rotation risk (-1)
Model O2.5: 37.7% | Market O2.5: 59.9% | Gap: -22.15% | BTTS: 41.2%
The standout analytical fixture of the gameweek. The model sees only 37.7% Over probability — meaning it strongly favours Under 2.5 — while the market implies 59.9%, a staggering -22.15pp divergence. Both teams carry identical label sets: <Defensive Block> and <Cold Attack>.
Both sides are structurally low-scoring and both are currently below their season averages in attack. The context overlay adds Forest’s European rotation risk on top of an already bleak attacking picture. The BTTS figure of 41.2% is the lowest in the gameweek — the model doesn’t expect both teams to score, let alone the game to produce three or more goals.
The market at 59.9% is pricing this as a coinflip on goals. The model disagrees emphatically. This gap sits far outside our validated band so we do not flag it — but of all the fixtures this week, this is the one where the model and market are furthest apart, and the label evidence is entirely consistent with the model’s view.
Everton vs Man City Monday 4 May
Everton <In Form Attack>
Man City <High Scorer> · <In Form Attack>
Model O2.5: 79.2% | Market O2.5: 62.5% | Gap: +16.74% | BTTS: 66.0%
The largest positive gap of the gameweek and a fixture that on paper looks like a clear high-scoring game. Man City arrive as <High Scorer> and <In Form Attack> — one of the strongest attacking label combinations in the division — visiting an Everton side also showing <In Form Attack>.
The model’s 79.2% Over probability reflects the quality of City’s attack against an Everton side that has been creating chances. The market at 62.5% is more cautious — perhaps reflecting Everton’s defensive record or City’s tendency to manage games once ahead. Gap at +16.74pp is well above our cap and we do not flag it, but the underlying data strongly supports a goal-heavy fixture.
Gameweek 35 — Selections
No confirmed selections this gameweek. Every gap either sits above the 6pp cap or below the 4pp floor. Two consecutive gameweeks without a flag is the model working correctly — we do not force selections when the validated edge is not present.
The most analytically interesting fixture this week is Chelsea vs Nottingham Forest. Both teams <Defensive Block> and <Cold Attack>, Forest with European rotation risk applied, model at 37.7% Over against a market pricing 59.9%. The data is telling a consistent and emphatic Under story. We cannot act on a gap this large within our validated framework — but it is worth monitoring whether the market moves before Monday.
CLV tracking: 3 selections logged, 2 wins, 66.7% win rate, 66.7% CLV+ rate. Sample too small to draw conclusions — track 20 plus selections before reading into these numbers.
Valuefinder now has a web app! Find the full football model results at valuefinder.app/football