The New Saintsvs
Trefelin
First halves carry fewer goals, and the model's half-time grid reflects it. The New Saints won their latest game (WWD). Probability 73%.
Match modelFriday 20 November: 165 fixtures clear the half-time threshold, with a mean probability of 44%. Highest reading: The New Saints vs Trefelin, The New Saints ahead at half-time at 73%.
Home ahead, level or away ahead at the break, using the first-half share of expected goals. Sorted by probability; tap a match for its score grid.
First halves carry fewer goals, and the model's half-time grid reflects it. The New Saints won their latest game (WWD). Probability 73%.
Match modelFirst halves carry fewer goals, and the model's half-time grid reflects it. Bayern München come in off a win (WWD). Probability 64%.
Match modelHalf-time grid built from 44% of the 3.39 total xG. Triglav come in off a win (WWL); Dren Vrhnika: DLL (0-1-2). Probability 60%.
Match modelFirst halves carry fewer goals, and the model's half-time grid reflects it. Probability 59%.
Match modelFriday 20 November: 165 selections, mean probability 44%. Strongest upcoming reading: The New Saints vs Trefelin (Premier League) – The New Saints ahead at half-time, 73%. Busiest competitions: Netherlands Eerste Divisie (9), Algeria Ligue 2 (8), Wales Premier League (8).
Half-time result is a three-way market on the score at the break. The model assumes 44% of the expected goals arrive in the first half.
An X pick wins on any level half-time score, whatever happens afterwards.
With fewer goals before the break, the draw is the most likely half-time result in most games; a home-lead pick needs a strong favourite with a high xG.
Expected goals for each side come from xGaura's attack and defence ratings, a statistical xG feed and the goal line implied by bookmaker prices. A Poisson model turns those two numbers into the probability of every scoreline, and each market is a sum of cells from that one grid – so the 1X2, goals and BTTS figures for a match always agree. Fair odds are 1 ÷ probability; a price above them is flagged as value. See the methodology or the full grids on mathematical models.
Want combinations instead? Open the accumulators, the pick of the day or the market comparison.
The half-time result market settles on the score at the interval: the hosts in front, the scores tied, or the visitors in front. It is a separate betting contest from the full-time result, with its own probabilities and its own quirks – most notably that a level score is far more common at half-time than at full time.
Across football, a little under half of all goals – around 44% – are scored before the break. The model applies that share to each team's full-match expected goals to get first-half figures. A home side projected at 1.8 goals over 90 minutes becomes about 0.79 for the first half; an away side at 1.0 becomes 0.44. A separate Poisson grid built from those smaller figures produces the half-time probabilities.
With expected goals that low, 0–0 is often the single most likely half-time score. In the example above the chance of a goalless first half is e^−(0.79 + 0.44) ≈ 29%, and adding 1–1 lifts the half-time draw to roughly 38–42%. That is why the half-time draw is frequently the selected outcome on this page, even in matches with a clear full-time favourite.
Each match's strongest half-time reading is listed with its probability, fair odds and expected value. Favourites rarely exceed 55% at half-time unless the gap in expected goals is very large. Because the market is less popular than full-time 1X2, prices can be less efficient, and value sometimes appears where the bookmaker has simply scaled the full-time odds rather than modelling the first half separately.
The ratings capture overall strength, not timing patterns, so treat half-time readings with a little more caution than full-time ones.
The first-half goals page uses the same first-half expected goals for over/under lines before the break. The half-time/full-time page combines both halves into nine outcomes. For the full-match result, see 1X2 predictions.
The bet settles on the score at the half-time whistle, including first-half stoppage time. The market's record is on the leaderboard. Over 18s only.
A match is projected at 1.6 home and 1.2 away for 90 minutes. First-half figures become roughly 0.70 and 0.53. The grid then gives the home side about 35% to lead at half-time, the away side about 24%, and a level score about 41%. Compare that with the full-time picture – roughly 47%, 25% and 28% – and you can see why the half-time draw is so often the strongest reading.
Because first-half expected goals are small, a single fast start or slow start changes this market more than most. If you know a team habitually presses hard in the opening twenty minutes, treat that as information the league-wide average does not contain, and weigh it alongside the model's figure rather than ignoring it.
Yes – the result at the half-time whistle counts.
Teams start cautiously and fewer goals are scored before the break.
Each side's expected goals come from attack and defence ratings plus market prices. A Poisson grid turns them into the chance of every score, and the half-time figure is the sum of the matching cells. The list is sorted by that probability.
How often the model expects the pick to win. 70% means about seven wins and three losses in ten matches with the same numbers.
Yes – every pick is graded on the 90-minute score and kept on the Yesterday tab and the track record.
Enter your email once to unlock the full score grids, stake guides and the 90-day leaderboard on this device. No password, no payment. We send at most one weekly model report; unsubscribe any time.
18+ · See our privacy policy.