Tell apart a team that is good and one that has been fortunate.
A team is fourth, having won six matches by a single goal and lost none of them narrowly. What should you expect next?
Yes.Close results are close to coin tosses. A team whose position rests on winning nearly all of them has been fortunate, and fortune is the one component that does not carry forward.
Not quite.The ability to win close matches has been looked for repeatedly and is very hard to find. What repeats is creating more chances than the opposition.
Not quite.This is one of the more predictable patterns there is: extreme results in the noisy component tend to come back toward the middle.
Results are made of two parts: how well you played, and how the close moments fell. Only the first one carries into next month.
Two teams, same points. Slide to see what produced them.
Team ATeam BSix months later
Good chances created2.1 per match
Good chances conceded1.2 per match
Team ACreated far more chances than it conceded, and lost several matches by a single goal.
Good chances created1.1 per match
Good chances conceded1.9 per match
Team BWas outplayed most weeks and won a string of tight games.
Team A, next 19 games34 points
Team B, next 19 games19 points
Six months laterA climbed and B fell, with no change in manager or squad at either club.
Both teams had the same points. What told you they were not the same?
Yes.A season has 38 results and thousands of chances. Counting the underlying events gets you a much less noisy measurement of the same thing, which is why every serious club now does it.
Not quite.The divergence is predictable, and it is predicted from the chance counts before it happens.
Not quite.Neither manager changed, and both teams moved in the direction the chance counts pointed.
Move the control to see what changes.
Which of these measurements are stable enough to predict the next few months?
Chances created and conceded per match.
Share of possession in the opposition half.
Points from the last five matches.
Percentage of shots that went in this month.
Record over the past two seasons.
Yes.The stable measures count many events; the unstable ones count few, or count a rate that bounces around. Shooting percentage over a month is the classic trap — it feels like form and behaves like noise.
A team scored on 22% of its good chances this season, against a long-run average of 13%. If it created 100 chances next season, how many fewer goals would the average rate give?
goals
Yes.22 minus 13. Nine goals is worth roughly five or six points — a fall of several places, produced entirely by a rate returning to normal.
Put in order how to read a league table honestly.
Tap them in order — first to last.
Points→Games played→Chances, not goals→Skill vs fortune
Four steps between a table and a judgement.Yes.Each step strips away one layer of noise from the same underlying question: how good is this team? The table answers it last, not first.
What is the honest summary of what a season measures?
Yes.It covers all of it: the noise in single matches, the length needed for a table to mean anything, and why clubs count chances. The table is evidence, with a known error bar.
Not quite.Every lesson here has been about the gap between the result and the quality behind it.
Not quite.Over 38 matches, quality dominates at the extremes. Luck decides the middle, which is where most of the arguing happens.
Lesson complete
Only the part of a result you earned carries into next month.