The ledger · every number checked

When we say a number, how close does it land?

Every projection is graded against the points that actually happened — the good weeks and the bad ones, because this page is generated from the grading, not from marketing. The short version: When Onside says 4.2 points, players score 4.3. When FPL says 4.1, they score 3.2. Closer to the real score than FPL in 120 of 120 gameweeks. When Onside says a player will very likely (80%) score inside a range, he does 80% of the time.
of scores inside our 80% range
80%
said 80%, happened 80%
gameweeks closer than FPL
120 of 120
blind test, four seasons
we say → players score
4.2 → 4.3
FPL says 4.1, they score 3.2
predictions graded
90,150
the engine had never seen them

For the stats people · How we tested it: each of four seasons (2023-24 to 2026-27) was predicted using only what the engine learned from the other three, so it never saw the answers. 90,150 predictions over 120 gameweeks. "Says X, scores X" takes every player given about the same number and checks what they scored on average. "FPL" is FPL's own predicted points, worked out as they stood at each past deadline. "Best XI" is the eleven players rated highest each gameweek, with no budget or formation limit, the same rule for both. The 2025/26 season below is one of those four, shown week by week.

Week by week: whose number was closer?

25/26 · 28,913 GRADED
0.81.11.4ONSIDEFPL XPGW2–GW38 · LOWER IS BETTER

Two lines, one question: whose guess landed closer to what actually happened? The engine was closer than FPL’s own projection in 97% of gameweeks, closer than “just follow recent form” in 100%, and closer than a season-average guess in 100%. The weeks it lost are on the chart too — that’s the point of the page.

How close we get, position by position

Goalkeepers0.59
Defenders1.12
Midfielders0.94
Forwards1.06

The bar is our typical miss in points — shorter is better. Keepers are the most predictable players in the game; defenders the least, because their returns arrive in lumps of clean sheets.

When a player actually scored, who saw it coming?

CLOSER IS BETTER

One average buries the question a manager actually cares about: when a player returned, did the model see it coming? So here is the same record cut by what happened on the pitch. Misses are naturally bigger on haul weeks for every model in existence — the column to read is the last one: how much closer we were than FPL’s own projection, in points.

Onside's miss against FPL's own projection, in points (RMSE), cut by what the player actually scored. Positive edge means Onside's number was closer.
WHAT HAPPENEDPLAYERSOUR MISSFPL'S MISSWE'RE CLOSER BY
Steady returns 3–4 pts
our biggest edge — we see the quiet contributors coming
1,6451.342.17+0.82 pts
Blanks played, 2 pts or fewer
we flag the weeks a player does nothing
7,0641.572.23+0.66 pts
Big hauls 5+ pts
the hardest thing in FPL to call for any model; closer here too
2,4485.465.67+0.20 pts
No-shows didn't play
minutes are modelled, not assumed, so the players who sit out cost less
17,7560.750.87+0.13 pts

Same 28,913 graded predictions as above, cut four ways instead of averaged once. Positive means our number was the closer one.

For the stats people · “miss” here is RMSE (root mean squared error), which weights the big misses hardest — the strictest grading available.

V7 ENGINE · PRE-SEASON · AS OF 21 NOV · ONSIDEARENA.COM