Math & ModelsUpdated 28 Jul 20264 min read
Football Form: What W-D-L Actually Tells You
Form is five results with the opponents removed. It is the most prominent column on most tables and the one carrying the least information.
Every league table ends with a strip of five letters, and readers reach for it first. It is worth knowing exactly what it does and does not contain.
Which way does it read?#
Left to right, oldest to newest. The letter on the right is the most recent match.
This sounds too obvious to state, and it is not. The underlying data almost always arrives newest-first from the provider, and turning it around is a step a site can simply forget. Ours did:
Both of our data pipelines store form newest-first while every surface documented the opposite. We found it by replaying 900 rows against the actual fixture list, with palindromes excluded so a string could not satisfy both readings: 750 were newest-first against 4 newest-last, across 188 competitions. The tooltip had been naming the wrong match — calling last weekend's result "5 matches ago" — and a display helper was taking the five oldest results on the rows where the provider sends more than five.
If a site does not tell you which direction its form runs, one check settles it: find a team that lost last weekend and see which end the L is on.
What the five letters leave out#
The opponents. This is the whole problem. Form treats a win over the leaders and a win over the bottom club as the same letter. Fixture lists are not balanced over any five-match window, so two teams with identical strings can have played completely different football.
The margin. WWWWW from five 1–0 wins and WWWWW from five 4–0 wins are the same string. Goal difference keeps that information; form discards it.
Home and away. Three home matches against two is a meaningful difference over a five-match window — see home advantage for how much.
Whether it was deserved. A team can win five while being outplayed in most of them. That is exactly what expected points measures, and where the two disagree is more interesting than either alone.
The sample-size problem, concretely#
Five matches is not a sample from which to conclude much. Consider a genuinely average team — one that wins, draws and loses at the league's base rates. Runs happen to it constantly:
A run is what a sequence of coin flips looks like. It becomes evidence only when it is long enough to be unlikely, and five matches is nowhere near that threshold. The general treatment is in sample size; the short version is that almost every "in form" and "out of form" narrative is a description of noise given a name.
Where form does carry something#
It is not worthless. Two cases where it is genuinely informative:
- When it reflects something structural. A team that lost its first-choice goalkeeper and centre-back four matches ago has a form string that is describing a real, persisting change. The letters are a symptom; the squad is the cause. The information is in the cause.
- As a prompt, not a conclusion. A striking run is a reason to look at why — fixtures, injuries, a manager change, the underlying numbers. It is not a reason on its own.
Form is the most visible statistic in football, which makes it the least likely to be mispriced. By the time a run is obvious enough for you to notice, it is in the odds — and because the run is more salient than it is meaningful, the price has often moved further than the evidence justifies. That argues for fading obvious form more often than following it, though not reliably enough to be a strategy.
Reading ours#
The strip beside each club on our tables shows the competition's own five-match form, oldest on the left. Green is a win, grey a draw, red a loss. Letters the provider sends that are none of those are kept rather than dropped — MLS Next Pro settles draws by shootout and sends O for it, on 76 rows in our data.
We show it because a table without it looks incomplete, and we put it last because that is roughly its rank among the columns worth your attention.
Related
Frequently asked questions
Which end of a form string is the most recent match?
On Tofiko the newest result is on the RIGHT — the strip reads left to right, oldest to newest, like a timeline. This is worth checking on any site you read, because the underlying data usually arrives newest-first and the display does not always turn it around. Ours did not for a while: both of our data pipelines store form newest-first while every surface described the opposite.
How much does form predict the next result?
Much less than its prominence suggests. Five matches is a very small sample, and the five results carry no information about who the opponents were. A run of WWWWW against the bottom five and a run of LLLLL against the top five can describe the same team.
Why does a team's form look different on another site?
Because there is no standard for what counts. Some sites include cup matches, some only league fixtures, some show six or ten rather than five. Ours shows the competition's own form string for that competition, which is why it matches the table it sits next to.
What does the O in a form string mean?
A shootout result. MLS Next Pro decides drawn matches by penalty shootout and the provider sends O for it. We keep letters we do not recognise rather than dropping them — dropping one would silently shorten a club's form and shift every result in the strip by one place.
Is a losing run a reason to bet against a team?
Not by itself, and often the opposite. If a market has also noticed the run, the price has already moved, and it may have moved further than the evidence justifies — five matches is not much evidence. Betting on form is usually betting on information the market priced before you saw it.