28 August 2026

How Many Games Should a League Season Actually Be?

Season length is the quietest decision a league makes and the one that decides who wins. Eight games crowns whoever ran hot in August. Thirty games crowns whoever turned up in February.

Most leagues never really decide it. Someone picks a number in month one, the calendar fills in around it, and three years later the league still runs 26 game seasons nobody chose.

What season length actually controls

Three things feed your standings: how well people finish, how often they show up, and how the night falls. Length changes the mix. Short seasons give variance a big share. Long seasons grind it down until attendance and steady finishing are the whole story. Neither is wrong, but they hand out different trophies.

Short seasons: 8 to 12 games

You get readable standings, more champions per year, and a low barrier for late joiners. Casual players will commit to eight weeks when they will not commit to eight months.

You pay for it in variance. Points scale with field size, so a win in a 30 player game pays 10 x 30 = 300 while a win in a 12 player game pays 10 x 12 = 120. Say a player scores in half their games, and a typical scoring night is fourth in a 20 player field, worth 6 x 20 = 120. Over 8 games that is 4 scoring nights. Make one of them a standout, a win in a 30 player field worth 300, and the season total is 3 x 120 + 300 = 660. Run the same player over 30 games: 15 scoring nights, 14 ordinary and the same standout, is 14 x 120 + 300 = 1,980. That one night is 45 percent of the short season and 15 percent of the long one.

Long seasons: 20 to 40 games

You get standings that are a real record of the year. Nobody wins 30 games on a hot streak, and the middle of the table sorts itself honestly.

You pay for it when the race ends early. Suppose after 20 of 30 games the leader has 2,400 points and fifth place has 1,600. The chaser does not need 800 more points, they need to outscore the leader by 80 a week for 10 straight weeks while the leader keeps finishing high. That effectively requires the leader to stop showing up, and once the room works this out the season is over in everything but name.

A second effect catches organizers out. When mid table players decide they cannot win, they stop coming and fields shrink, which deflates every remaining game. The nights a chaser needs to be worth 300 are now worth 120, so the lead is safer than the raw gap looks.

Best N of M, and the arithmetic it really does

The standard fix for a long season is to count only a player's best N scores out of M games, so that someone who misses four Thursdays for work or family is not out of the race by March. Here is what it actually does, in a 16 game season where every field is 20 players.

• Dana plays all 16 and finishes fifth every time: 5 x 20 = 100 a night, so 16 x 100 = 1,600.
• Marcus works weekends and plays 12. He finishes fourth every time: 6 x 20 = 120 a night, so 12 x 120 = 1,440.

Counting every game, Dana wins by 160. Switch to best 12 of 16 and Dana keeps only her twelve best nights, 12 x 100 = 1,200, while all of Marcus's 12 games still count at 1,440. Marcus wins by 240. Nothing changed except the rule.

That is the honest trade: best N of M converts a reward for attendance into a reward for per game finishing. Set N too low and the failure mode is obvious. Under best 8 of 16, a player who shows up 8 times and averages third in a 20 player field scores 8 x 7 x 20 = 1,120 and beats Dana's capped 800. Half a season wins it.

A workable rule: drop the number of games a reasonable person misses for life, not the number a part timer misses by choice. Usually 2 to 4. Best 14 of 16. Best 20 of 24. Best 26 of 30.

One caveat. Under the default table only the top nine score, so many of a regular's dropped nights are already zeros, and dropping a zero changes nothing. The rule only bites for players who score in more than N games, as Dana does above. If most of your field finishes outside the scoring places most weeks, lengthen the table before you reach for a drop rule: a drop rule needs something to drop.

Season length versus the calendar you have

Pick the game count second. Count the playable dates first. A weekly league has 52 slots on paper, but Christmas, New Year, Thanksgiving, the summer holiday weekend, and a couple of venue closures take a bite. Call it 46 usable nights: two seasons of 23, three of about 15, or four of about 11.

Quarterly seasons of 8 to 12 games are the easiest to explain to a room and to sell to a venue, because the boundary lands near a natural gap anyway. If you prefer two long seasons, put the break where attendance already dips, usually high summer or right after the holidays. That turns your worst weeks into a reset instead of a slow decline. And never put a season's last night on a holiday weekend.

Close the season with a finale, not a spreadsheet

A season that ends with the last ordinary game feels like it stopped rather than finished. A finale gives the calendar an end point, and it gives the players in eighth through fifteenth a reason to keep turning up in the closing weeks, which is the stretch where a long season loses them.

As far as the standings are concerned the finale is a game night like any other: everyone who qualifies plays, you enter the finishing order, and it scores the way every other night scores.

Publish two things in game one. The qualification line: top 16, top 20, anyone who played six games. And which title the night decides. The safe split is to award the season on season points and the finale on the night itself, so the leader keeps what the season earned and the room still has a night worth winning. A league that awards one thing twice will spend the winter arguing about which one counted.

When to reset, and how not to erase anyone

Organizers hesitate to reset because it feels like deleting five years of somebody's league record. The regular since day one does not want to sit at zero next to someone who joined in April.

That is a software problem, not a scheduling one. Season standings reset on schedule and drive the current race, while player profiles carry career stats and head to head records that never reset, alongside the trophies the season handed out. Nobody loses their history when a season starts clean, so a reset stops being a loss and becomes the point.

A starting point

With no strong opinion yet, run 10 to 14 games a season, three seasons a year, drop the two lowest scores, and close each one with a finale for whoever qualifies. That survives holidays, keeps the standings live into the last month, and hands out three champions a year instead of one.

Run this on your own league

Paste a season of finishing orders in, then delete half the games and look again. Season length is a scoring decision, and the points simulator is where you find out what yours is doing.

Back to every article