How to Project a Season Record With Simple Math

A season projection turns opinions into a forecast. Instead of saying a team looks good, you can say it has a 62 percent chance to win eight games. I build these forecasts by hand, and the math is simple addition.

I will show each step with numbers you can copy. You will give every game a win chance, add them up, and see the full range of final records. After that, you can update the forecast each week as results come in.

Key takeaways

• Give every game its own win chance.
• Add the chances to get expected wins.
• Study the full range of outcomes, because the most likely record is not the only one.
• Update the numbers weekly, then grade your forecast at season's end.

What a projection is, and what it is not

A projection is a set of win chances, one per game, combined into a forecast for the year. It does not predict that a team wins exactly eight games. It says how likely each total is. That difference matters, because a single number hides the risk in both directions.

Think of weather. A 70 percent chance of rain does not promise rain. It tells you to carry an umbrella. A season projection works the same way. It tells you what to expect and how wrong you could be.

Step 1: give every game a win chance

Start with a number from 0 to 100 for each game. Three inputs help most.

• Team strength. Compare yards per play and turnover margin for both sides. My yards per play guide shows how.
• Home or road. A home game usually earns a higher number than the same game on the road.
• The schedule. My schedule guide explains how to judge the opponent.

Early in the year, you may not have much data on later opponents. In that case, use a tier. I rate each opponent as a clear underdog, a slight underdog, an even game, a slight favorite, or a clear favorite. Then I attach a number to each tier.

Those tier numbers are my own choices, not facts. Change them to match your view. The point is to make every opinion visible, so you can fix it later.

Step 2: add the chances

The sum of your win chances equals the expected win total. The math works because the average result of each game equals its win chance. Add the averages and you get the average of the season.

The team in this example is expected to win 7.2 games. No team wins 7.2 games, of course. The figure is an average, and the real total will land above or below it.

Step 3: find the range of outcomes

An average hides the spread of results. To see the range, you need the chance of each possible total. A small example shows how.

Take three games with win chances of 70, 50, and 30 percent. The chance of zero wins is 0.30 × 0.50 × 0.70, which equals 0.105. The chance of three wins is 0.70 × 0.50 × 0.30, which also equals 0.105. For exactly one win, add three paths. Win only the first game (0.245), win only the second (0.105), or win only the third (0.045). The sum is 0.395. The remaining chance, 0.395, belongs to two wins.

The expected total is 1.5 wins, but 1.5 is not even a possible record. The chance of two or more wins is 79 percent. This is the kind of statement a projection lets you make.

The adding-one-game rule

You do not need to list every path for a long schedule. A short rule lets you add one game at a time. For each total k, the new chance equals the old chance of k times the chance of losing the new game, plus the old chance of k minus 1 times the chance of winning it.

Add a fourth game at 60 percent to the table above. The chance of zero wins becomes 0.105 × 0.40, or 4.2 percent. One win becomes 0.395 × 0.40 plus 0.105 × 0.60, which is 22.1 percent. Two wins becomes 0.395 × 0.40 plus 0.395 × 0.60, or 39.5 percent. Three wins becomes 0.105 × 0.40 plus 0.395 × 0.60, which is 27.9 percent. Four wins is 0.105 × 0.60, or 6.3 percent. Those five numbers sum to 100 percent, and the expected total is 2.1 wins.

A spreadsheet or my season projector repeats this rule for each game on the schedule. You enter the chances and read the output.

How wide is the range?

A quick measure of spread is the standard deviation. Multiply each game's win chance by its loss chance, add the products, and take the square root. For the 12-game example, four games at 80 percent give 4 × 0.16, or 0.64. Four at 60 percent give 4 × 0.24, or 0.96. Four at 40 percent also give 0.96. The sum is 2.56, and the square root is 1.6.

That means the team's total will often land within about 1.6 wins of 7.2. In plain terms, a range from 5.6 to 8.8 covers about two-thirds of the likely outcomes. A 6-6 or 9-3 finish would not shock anyone.

Step 4: set a goal and read the chances

Pick a target, such as eight wins. Add the chances of every total at or above that line. For the 12-game team, the chance of reaching eight or more wins comes out to 43.2 percent. That is lower than a coin flip, even though the average of 7.2 sounds close to eight.

Read the table as a menu of outcomes. The most likely single record is 7-5, yet it happens less than one time in four. A double-digit win season is possible, but it sits near 7 percent.

Goals tell a story that averages cannot. A team with 7.2 expected wins has a decent shot at six wins and a fair shot at eight. Use the goal that matters to you, whether it is a winning record, a bowl bid, or a place in the conference race.

A conference-only projection

The SEC race rests on nine league games in 2026. So build a second projection that counts only those nine. Say a team's league win chances are 85, 80, 70, 65, 60, 55, 45, 40, and 30 percent. They add up to 5.3 expected league wins. The standard deviation is about 1.4 wins.

That 5.3 means the team most likely finishes 5-4 or 6-3 in conference play. The two spots in the title game belong to the top two teams in the overall standings. My standings guide explains how win percentage and tiebreakers decide those spots. A 6-3 league mark is a .667 win percentage, and a 7-2 mark is .778.

Update the projection every week

A projection should change as the season moves. After a game, replace its win chance with a 1 for a win or a 0 for a loss. Then adjust the later games for new information, such as an injury, a coaching move, or a surprise result.

1. Record the final result.
2. Replace that game's chance with 1 or 0.
3. Re-rate the remaining opponents using fresh spreads or notes.
4. Recompute the expected wins and the range.
5. Write one sentence on what changed and why.

The last step builds the habit that separates good forecasters from lucky ones. You will see which beliefs moved and which stayed.

Where the method breaks

This approach assumes each game stands alone. Real seasons do not behave that way. If a starting quarterback gets hurt, every remaining game changes at once. The wins and losses become tied together, so the true range is wider than the simple math shows.

Early-season chances are also soft. A rating based on last year's team can miss a coaching change or a transfer class. Treat September projections as rough sketches and November projections as sharper tools.

Finally, remember that the win chances are your inputs. A neat table cannot repair a bad input. If you rate your favorite team too high in every game, the projection will tell you what you want to hear. That is a trap.

Check your own accuracy

You can grade yourself with a simple score. For each game, subtract the result (1 for a win, 0 for a loss) from your win chance and square the difference. Average those squares across games. The measure is called the Brier score, and lower is better.

Say you gave 80 percent to a game your team won, 60 percent to one it lost, and 70 percent to one it won. The squared errors are 0.04, 0.36, and 0.09. Their average is 0.163. As a reference, someone who rates every game at 50 percent scores 0.25. Beating 0.25 means your ratings added information.

How to use the season projector

1. Open the projector on my SEC football analysis tool page.
2. Pick the number of games.
3. Enter a win chance for each game.
4. Choose a goal, such as eight wins.
5. Read the expected wins, the most likely record, and the chance of reaching your goal.

Try two versions of the same schedule. Use your honest ratings first, then a cautious set that lowers each chance by five points. The difference shows how sensitive the result is to your assumptions.

Run the same drill each week. By October, you will see how much a single result can move the forecast.

Build an SEC projection in two blocks

A 2026 SEC schedule has 12 games. Nine are league games, and three are nonconference games. Set the three nonconference chances first. Then rate each league opponent one by one.

Two blocks keep an easy start from hiding a hard finish. They also match the title race, since only the top two teams in the overall standings reach the Dec. 5 game in Atlanta. A league-only projection tells you more about that race than a full-year total.

My take

I never quote expected wins alone. One number hides the spread of outcomes, and the spread is where the story lives. I would rather see the full range, because a team with the same average can have a very different chance of a great year.

The numbers I ran

I built one made-up 2026 schedule. Three nonconference games each get a 92 percent chance. The nine league games get 80, 72, 65, 58, 52, 45, 40, 33, and 25 percent. I then worked out every possible record.

The team averages about 7.5 wins, yet its chance at 9 wins is well under half. That gap is why I look at the range.

Common mistakes

• Treating the expected total as a promise.
• Rating every game above 50 percent for a team you like.
• Forgetting that three 70 percent games do not make a 70 percent chance to win all three. That chance is 34.3 percent.
• Skipping the weekly update.
• Ignoring injuries that hit several games at once.

Often asked questions

How do I project a college football team's record?

Give each game a win chance, then add the chances. The total is the expected number of wins.

What does expected wins mean?

It is the average number of wins over many imagined seasons. It does not have to be a whole number.

How do I find the chance of reaching eight wins?

Compute the chance of each possible win total, then add the chances for eight wins and higher.

Where do I get win chances for each game?

Use tiers you set yourself, and adjust for schedule strength and home field. My guides explain both.

What is the standard deviation of wins?

Multiply each game's win chance by its loss chance and add the products. The square root of the sum is the standard deviation.

How often should I update a projection?

Update it every week. Replace played games with 1 or 0 and re-rate the games that remain.

Why is my projection wider than expected?

Injuries and form changes link games together. That makes real outcomes more spread out than the simple math shows.

What is a Brier score?

It is the average squared gap between your win chances and the actual results. A lower score means better forecasts.

The takeaway

A good projection is honest arithmetic. Rate each game, add the chances, and study the range. Update the numbers each week and grade yourself at the end. When your numbers are ready, run them through the season projector on the tool page.