Sports projections made simple.

A short explainer for anyone visiting the site for the first time — no sports-betting background required.

What is GameTimePicks?

GameTimePicks is an educational sports analytics project. For each sport we cover, a model publishes its forecast before the game starts — who is favoured, likely scores, player numbers — and every forecast is graded against the official result afterwards, losses included, so the track record stays honest. Where a sportsbook price exists it is shown beside ours for comparison. It's a research lab — not betting advice.

How projections work

For each event on the slate we pull the inputs the sport's own model declares — game logs, season rates and home/away context for MLB player markets; team-strength ratings for NFL; an Elo team rating and a player simulation for NBA; a goal-rate match model for the Premier League, applied unchanged to Ligue 1; tracked fight history for UFC — and run a deterministic simulation. Where a sportsbook line exists it is shown for context, never as an input. We never invent inputs; if a player's log or a fighter's history is missing, the read is suppressed and the page says so.

How to read a projection

Line · the number the bookmaker is offering Over/Under.
Projection · the model's estimate for that player tonight.
Gap / Edge · Model % minus Market % — how much our model disagrees with the market. It compares two probabilities, not the projection and the line — so a projection below the line can still show a positive gap, when the model thinks the Over is likelier than the price implies. Market Guide carries the full definition and its caution.
Side · Over if the projection is above the line, Under if it's below.

What the category labels mean

Each projection falls into one of three categories, based purely on how far the model’s number sat from the sportsbook’s. The category describes how a row was produced. It does not rank quality, and it does not affect the order anything is shown in.

Category A · model and market differed by 5pp or more. Settled at 49.4%.
Category B · differed by 2.5–5pp. Settled at 50.2%.
Category C · differed by under 2.5pp, or the row was anomaly-flagged. Settled at 50.9%.

Those rates are measured over 41,007 settled MLB player-prop leans (as of 2026-09-12), and they run in the opposite direction to what the old labels implied. A larger disagreement with the market has historically settled worse, not better — which is why these are neutral letters now, and why no category is promoted above another anywhere on the site.

Why results matter

A track record is the only honest claim a projection site can make. We publish every settled projection on the Results page — wins, losses, and pushes — and grade after the final box score. Pushes are excluded from the hit-rate denominator; pending games never count as losses. The deep-dive technical breakdown lives at /results/model-audit.

Responsible use

This is research and analytics, not betting advice. Don't risk money you can't afford to lose. If gambling is becoming a problem, the resources on the Responsible Use page can help.

What's coming next

Wider market coverage in each sport we cover, as each market clears its own evaluation bar — never before.

History · the model watchlist as it read on May 24, 2026An archived snapshot from the NBA/MLB era of the site. It does not update; current model states are on System Status and each sport hub.

Model watchlist — archived snapshot (May 24, 2026)

An archived read from the NBA/MLB era of the site, kept for the record. For current model states, see System Status and each sport hub's coverage table — those derive from live receipts; this snapshot does not update.

NBA rebounds — the strongest cohort on record. The model has stable signal on REB projections.
NBA points + assists — barely above coin flip on a large sample. We surface these projections but treat them as watch-list calls, not high-confidence reads.
MLB strikeouts — smallest sample of any market we cover and below coin flip so far. The variance profile of pitcher hooks + manager decisions makes this an honest weak spot.
MLB confidence climbed back into "watch" on the May 22 settlement — as of that date High was 49.7% on 396 settled rows, Medium 50.4% on 141, Low 53.3% on 435. These are the May 22 figures, not current ones; the live rates are in the category captions above. Low is still the best MLB cohort, but only ONE rival now beats High by ≥1.5pp (was both before May 22). The calibration overlay auto-promotes MLB High from a "Needs more tracking" downgrade back to its raw label. The decision rule is pinned by tests — we only invert when ≥ 2 rivals beat by ≥ 1.5pp, so a single best-tier (Low) can't trigger inversion.
Monte Carlo internal validation — two-date check — May 22 looked promising (MC Strong 62.5%, Watch 65.2% on 311 joins). May 23 reverted to roughly coin flip (Strong 50.0%, Watch 29.4% on 287 joins). Two dates is too small to draw a conclusion. Across both: Strong 56.7% (17-13), Watch 50.0% (20-20), High-variance 50.2% (259-257). Promotion to production scoring is on hold until ≥ 5 dates show consistent separation.
Early tracking: curated picks vs saved parlays (2 days) — across the first 2 days of tracking, single-leg curated picks are 8-4 (66.7% on 12)while multi-leg saved parlays are 6-44 (12.0% on 50). MLB-only curated picks are 5-1 (83.3%). Two days and 62 settled picks are far too few to say either track is better, and a multi-leg slip loses more often by construction, since every leg has to land. Both tracks stay visible.
Curated rail prefers selectivity over volume — the homepage "Tonight's curated projections" rail picks up to six leans per slate by edge × calibration-adjusted confidence × market strength. Inverted (sport, tier) combos are excluded. Better to see six trustworthy reads than 300 of mixed quality. The picks are saved before games and graded after settlement — so the curated rail will eventually carry a real, auditable hit rate of its own.
Calibration is now derived from the live audit — the confidence overlay reads the model audit every render. When the nightly settle adds more data, labels adjust automatically. We fail closed: thin samples stay informational and inverted tiers are flagged. No category is ever promoted above another: on settled data the categories run in the opposite order to what their old names implied, so none of them earns priority.

Numbers are pulled from Results. Sample sizes are still small in absolute terms — anything you read here is a record, not a forecast. No 80%-accuracy claim is made anywhere on the site, and won't be until we run a real out-of-sample backtest.

Technical surfaces

Educational analytics · not betting advice