With a player sent off, a team wins about 9 fewer games in 100

The side that was left with ten men, because the referee sent a player off with a red card, won 19.5% of the games. The teams without a sending-off won 28.7%. Those teams are from other games in the same competition and were at the same score at the minute of the card.

That gap, about 9 wins in every 100 games with a sent-off player, shows the cost of the red card, which further down is also counted in points. This analysis measures that cost across thousands of games in leagues and cups from Europe (the Premier League among them), South America, Mexico and Japan. The cost is higher early in the game and lower near the end. The calculation does not prove that the card alone explains the gap, and the caveats come further down.

One case sums up the calculation: the game is tied, and the red card comes between the 16th and 30th minute. There were 1,017 such games, from 2004/05 to 2026. The ten-man side won 13.4% of them, and the teams without a sending-off at the same score, with eleven players until the end, won 33.3%. So the ten-man side won less than half as often as the teams with eleven.

How the calculation was made: the ten-man side against teams in the same situation

Only the first sending-off of each game counts, and 22,874 games had one. Another 359 games, in which both teams had a sending-off in the same minute, are left out. The first sending-off can be a straight red, not the result of two yellows (51.2% of the cases), or a second yellow (48.8%): a player who already has a yellow card and gets another in the same game is sent off. A red card shown to a substitute or a coach, or one overturned after a video review, does not count.

In 22,553 of those games, the Soccer Addicted archive, the portal's database of games, has teams in the same situation to compare with. Those are the games in the calculation, and the other 321 are left out.

For each of the 22,553 sending-offs, the comparison group gathers every team from the same competition and season in games with no sending-off at all. These are the teams in the same situation: at the minute of the card, they were at the same score and in the same place, home or away. What they did until the final whistle is the point of comparison. A goal scored in the same minute as the red card counts as coming before it.

Stoppage time, the extra minutes the referee adds at the end of each half, counts as minute 45 or 90. The calculation covers the 90 minutes, without extra time, the 30 additional minutes played in some cups.

Why not compare with all games? Because teams that are losing get more red cards. The sent-off team was losing in 39.6% of the cases, against 32.1% of the teams in the same season and minute, in games without a sending-off. Without this precaution, the cost would come out exaggerated, because part of the sent-off teams were already losing before the card.

Which competitions are in: 24, from the Premier League to the Copa Libertadores

There are 24 competitions, in 363 seasons and 122,901 games, with and without sending-offs, from 2004/05 to 2026. A season is only included if the goals and cards in the archive match the scores: the events must exist in almost every game, and the sum of the goals must match the final result in almost all of them. The FA Cup does not pass this test in any season. The UEFA Super Cup, with a single game per edition, has no other games in its season to serve as a comparison.

The 2026/27 season, which is under way, is left out, and so are the 2023/24 European competitions, because the archive does not have the first half of 2024. Also left out are 3 seasons with far fewer red cards than normal in their competitions, in which the archive may have lost cards.

The Brasileirão, the Brasileirão Série B and the 2026 Copa Libertadores are included with the games finished by 9 October 2026, the date of the calculation, because the season is still running.

What this tells us: each country has 1 to 3 competitions in the calculation, and Europe also comes in through the club cups, with 19 seasons of the Champions League and 16 of the Europa League. Inference: the DFB-Pokal, the Coppa Italia and the Coupe de France have 12, 7 and 4 seasons, and so the analysis shows the whole set, not one competition at a time.

The cost in league points: 0.29 per game, on average

To bring win, draw and defeat into a single number, the calculation uses league points: a win is worth 3 points, a draw 1 and a defeat none. Points per game is the average number of points a team earns in each match. The ten-man side earned 0.84 points per game, and the teams in the same situation 1.14. With more decimals, the difference is 0.29 points per game: the average cost of the red card.

There were also more defeats: the ten-man side lost 54.4% of the games, against 43.7% of the teams in the same situation. It drew 26.1%, against 27.7%.

In a tied game, with the red card between the 16th and 30th minute, the cost is higher: the ten-man side earned 0.65 points per game, and the teams in the same situation 1.31. The difference of 0.66 points per game is more than double the average of 0.29. The 95% confidence interval of the calculation, the range in which the number probably lies, runs from 0.60 to 0.71.

What this tells us: in a tied game with the red card between the 16th and 30th minute, the ten-man side lost 61.7% of the 1,017 games, against 35.5% of the teams in the same situation, and won 13.4%, against 33.3%. Inference: in every 100 such games, the red card came with about 20 fewer wins and 26 more defeats.

The earlier the red card, the higher the cost: from 0.73 to 0.10 points per game

The average hides big differences. The earlier the referee shows the card, the more points the team loses, and the later, the fewer. The table splits the game into six bands of 15 minutes and shows, for each one, how many games are in it, how the ten-man side did and the cost in points per game.

What this tells us: the cost goes from 0.73 points in the first 15 minutes to 0.10 in the last 15, and the confidence intervals of the six bands do not overlap. Inference: a red card in the first 15 minutes costs more than seven times as much as one at the end.

A late-game case is Eduardo Camavinga, of Real Madrid, sent off for a second yellow at 90+4, that is, 4 minutes past the 90th, when Real Madrid were losing 3-0 to Arsenal in the Champions League on 8 April 2025. That game is not in the calculation, because the 2024/25 Champions League does not pass the complete-data test. It only illustrates the kind of situation.

Early on, a red card weighs heavily, but it is not a certain defeat. In the first 15 minutes (843 games), the ten-man side lost 66.4%: 2 games in every 3, not all of them. One game in every 3 gave at least a point (33.6%), with 12.8% of wins and 20.8% of draws. The band from 16 to 30 minutes includes all scores, and so it differs from the case of the tied game.

At the end of the game, the cost is the lowest, but it is not zero. Almost 40% of the sending-offs come between the 76th and 90th minute (8,982 games), and there the cost is 0.10 points per game. With the game tied in that stretch (2,709 games), the ten-man side lost 16.4%, against 8.6% of the teams in the same situation.

Only when the team is already losing, and near the end, does the cost almost disappear. Between the 76th and 90th minute, it is 0.03 points per game (4,188 games). The reason is that a team that is losing at that point earns only 0.07 points per game, and the ten-man side, 0.04.

The score at the time of the red card: winning, tied or losing

With the team winning, the cost is 0.39 points per game, and with the game tied, 0.42. With the team losing, it is much lower, 0.11. A team that is losing already expects little: the team in the same situation earns only 0.21 points per game. The table shows the three scores.

What this tells us: with the team winning or the game tied, the red card costs from 0.39 to 0.42 points per game, and with the team losing, 0.11. Inference: the cost depends on how much the team still had to lose, and a team that was already losing had little.

The saying about the draw: with the game tied, the ten-man side draws less

The saying goes that, with ten, a team sits back and holds on to the draw. This can be tested with the score, because the archive does not measure what the team meant to do on the pitch. With the game tied at the time of the red card (8,528 games), 49.1% of the games ended in a draw, against 54.0% for the teams in the same situation, and 41.1% in defeat, against 23.7%.

Late in the game the draw falls less, but it still falls. Between the 76th and 90th minute, with the game tied, 79.2% of the games ended in a draw, against 83.2% for the teams in the same situation. Between the 16th and 30th minute, it was only 25.0%, against 31.2%.

The team that was winning (5,182 games) drew 21.7%, against 13.1%. The defeat was also bigger, 11.8% against 4.4%, and the win smaller, 66.6% against 82.5%. The draw grows less than the win falls.

On goals, both sides get worse. In the minutes left in the game (19,565 games with the red card before the 90th minute), the ten-man side scores 0.61 times what the teams in the same situation score (61 goals for every 100) and concedes 1.61 times what they concede (161 for every 100). The confidence intervals are from 0.59 to 0.62 and from 1.59 to 1.64.

Early or late, winning or losing: the cost in each situation

The table crosses the minute with the score and shows the cost of each situation. The highest figure shown is 0.79 points per game, with the red card between the 16th and 30th minute and the team winning, and the lowest is 0.03, between the 76th and 90th minute with the team losing. Two cells are left without a number because they have fewer than 100 games.

What this tells us: from the 16th to the 90th minute, the cost with the team winning or the game tied is higher than with the team losing, and it falls as the game goes on. Inference: the red card costs most when there is a lot of game ahead and there are still points to lose.

Playing at home does not protect, and the away team gets most of the red cards

At home and away, the cost is similar: 0.31 points per game for the home team (9,593 games) and 0.28 for the visitor, the team playing away (12,960). The home team earns more points with ten, 1.04 against 0.70, but it already earned more before: the teams in the same situation earn 1.36 at home and 0.97 away. What changes is who gets the card: the visitor received 57.5% of the first sending-offs (13,144 of 22,874).

The other side: the team with eleven wins 54.4% of the games

Against ten, the team with eleven is also part of the calculation. It won 54.4% of the games, against 43.7% of the teams in the same situation, and earned 0.31 more points per game. Almost half of the games did not end in a win for it. In the first 15 minutes, the team with eleven won 66.4%, and between the 76th and 90th, 50.3%.

Costing is not causing: what the calculation does not show

The calculation compares teams at the same score and minute, but not under the same pressure. The team that gets the red card may be the one that was already suffering on the pitch, or the one that committed the foul to prevent a goal. So the cost of 0.29 points per game is a difference between the ten-man side and the teams in the same situation, not proof that the card, on its own, took the points away.

Three more limits. The penalty that comes with the red card is inside the cost. The second yellow and the straight red are counted together, and the calculation does not measure the cost of each type. And the cost is an average: a game earns from 0 to 3 points, and none earns 0.29.

Changing the way of calculating hardly changes the result. Three ways of building the comparison group gave 0.29 in all three. Without the games in which the other team was also sent off later, it gave 0.31. In the leagues (19,778 games) and in the cups (2,775), it was 0.29 in both.

The minute of a card also varies between sources. In Atlético de Madrid 0-0 Manchester City, on 13 April 2022, ESPN puts a sending-off for Atlético at 90+1, and other sources differ by 1 to 2 minutes. The calculation uses bands of 15 minutes, not the exact minute. It describes 22,553 games that have already been played and does not predict the next one.

Next time the referee shows a red card

When you see a red card, look at the clock and the score. In the games in the archive, a red card between the 16th and 30th minute, with the game tied or the team winning, cost 0.66 to 0.79 points per game. Between the 76th and 90th, with the team losing, it cost 0.03. In my reading, this is the most useful calculation for anyone watching: the minute and the score show how much the red card cost in similar games, not what will happen in this one.

A clue the archive cannot yet prove: the cost in recent seasons

One clue is left unanswered, and the archive does not yet prove it. The cost seems higher in recent seasons, 0.33 points per game in seasons that started in 2020 or later, against 0.27 in those that started up to 2014. It appeared without a prior hypothesis, and so it is a clue, not a finding. The question for the next calculation is whether the red card is getting more expensive.