When 90 points was not enough: the anatomy of an LGT game
LGT 2026 in numbers | Part 5 of 6
Across 2,493 games, the scores reveal close finishes, high-scoring losses and a substantial gap between winning games and maximising points.
Scoring 90 points sounds like a good afternoon of Warhammer. At LGT 2026, it still ended in defeat 112 times across the first five rounds. There were also 563 games decided by ten points or fewer, including draws. Together, those figures explain why a final score needs an opponent's score beside it.
This analysis covers 2,493 games with two named players in rounds one to five. Unlike the faction win-rate comparisons, scoring figures include same-faction games. Each game contributes two player scores, giving us 4,986 scoring performances to examine.
What did a typical score look like?
The median player score was 85; the mean was 78.6. A quarter of scores were 65 or lower, while the upper-quarter threshold was 96. There were 840 scores of exactly 100. Of those, 836 were wins and four were drawn player scores, representing two 100-100 games.
The difference between the mean and median reflects the way lower scores pull down the average. A player could therefore score above the event mean without being above the middle score in the field. Neither measure, on its own, says whether that performance was enough to win.
High scores helped, but the opponent still mattered
There were 2,056 player appearances scoring at least 90. They produced 1,908 wins, 36 draws and 112 losses. The defeats included ninety scores in the 90-94 band and twenty-two in the 95-99 band. Even near-perfect scoring occasionally met something better.
The 80-89 band was particularly revealing: 422 wins, 34 draws and 426 losses. In this sample, a score in the eighties was almost equally likely to appear on the winning or losing side. That is a description of completed games, not a target score or a prediction to apply during play.
Close finishes were common, but not the majority
Of the 2,493 games, 563 ended within ten points, or 22.6%. That includes 43 draws, 275 games with a one-to-five-point margin and 245 with a six-to-ten-point margin. At the other end, 793 games finished more than forty points apart, or 31.8%. The event contained plenty of both narrow finishes and decisive score gaps.
Later rounds were closer on the median
The median margin fell from 32 points in round one to 22 in round five. That is a ten-point reduction, or 31.3% relative to the opening-round median. The mean margin also fell, from 33.8 to 27.6, although it was slightly lower in round four than round five.
It is consistent with later rounds producing more closely matched scores, but the exports cannot separate pairing effects from changes in who remained active or other round conditions. Recorded two-player games fell from 512 in round one to 477 in round five. We should not turn that descriptive trend into a claim that any particular mechanism caused it.
The score you allowed mattered too
Dark Angels averaged 84.6 points scored and 76.1 conceded. Orks averaged 82.8 scored and 72.8 conceded. Their average margins were therefore about +8.5 and +10.0 respectively. Orks scored fewer points on average than Dark Angels, yet had the larger average advantage over opponents.
That distinction is useful, but these totals cannot reveal how points were denied or whether a result came through primary objectives, secondaries or another scoring component. They describe the scoreline rather than the tactical route to it.
A perfect record did not require perfect points
The 32 players who went 5-0 accumulated between 363 and 500 points over those five rounds. Brett Smith, Feliks Bartkiewicz and Tim Lowes each recorded 500. Eventual champion Jean-Loup Genin reached 5-0 with 438.
A high total and an unbeaten record are different achievements. Points provide valuable context, but the range among the 5-0 players is a reminder that five successful games can look very different on the scoreboard.
About the data
Source: the LGT 2026 main-event placings and pairings exports. Scoring statistics use two-player rounds 1-5 games, including mirrors, and exclude one-sided entries. Close-game counts include draws. Points are exported battle points; no component-level scoring or in-game sequence is available.
Follow UKTC and the LGT over the next few days as we share the results, matchups and player stories from the event.
Comments