Follow Us ON

© 2017 Broom on the Warpath

Advertise Here Flag

What Does a 3-game Winning Streak Mean?


Sometimes I think there's something wrong with me. Chronic reality syndrome (CRS?) or something. The Wizards are riding the crest of a 3-game winning streak and fans are celebratory. Me? I'm saying, "So what?"

The problem, of course, is too much reading. It rots the brain, diminishes the ability to perform basic human functions like walking and talking, and can cause exhaustion and severe headaches the next morning. Wait, that's alcohol. Never mind.

As it happens, Dean Oliver (sorta the Bill James of basketball analytics) has a chapter about winning streaks in his excellent book, Basketball On Paper. So, what does a 3-game winning streak actually mean?

Not much, unfortunately.

In BoP, Oliver says that a .700 team should regularly have 3-game winning streaks. Over the course of an 82-game schedule, they should win all three of each independent 3-game stretch approximately 35% of the time.

A .300 team should have a 3-game winning streak about 3% of the time. But, that doesn't mean 3-game winning streaks are rare for a .300 team.

"Three percent of the time means that the .300 team will win 3 straight for 3% of the (independent) three-game sequences in a season," Oliver wrote. "Given that an NBA season is 82 games long, there are a lot of three-game sequences."

According to the handy table on page 70 of Oliver's book, a .300 team (24-25 wins) has an 89% chance of having a three-game winning streak during an 82-game season.

The Wizards are a .200 team at the moment (7-28). A .200 team has a 47% chance of winning three straight at some point during an 82-game schedule.

Aha! I hear from the rose-colored glasses crowd. You're not accounting for the fact that the Wizards have played just 35 games so far. So, how significant is a 3-game winning streak in a shorter schedule?

Umm, actually, not that significant. Oliver provides a second table on the same page -- this one for a 32-game schedule. I could build my own table for a 35-game schedule, but 32 is close enough.

According to Oliver's table, a .200 team (where the Wizards are now) has about a 21% chance of winning 3 in a row during a 32-game schedule. That percentage would rise a bit to a 35-game schedule -- probably to something in the neighborhood of 25%. A .300 team would have a 56% chance of winning three in a row in a 32-game schedule.

In other words, a 3-game winning streak does not mean the Wizards have turned a corner. It doesn't mean they've "gotten it" or figured out the "formula for winning."

At some point in an 82-game schedule, even a bad team is likely to play better than their opponents for a week or two.

Then, of course, we get to the issue of the team's lengthy losing streaks. Those games count too. Fans have a tendency of believing that the "real" team is the one that wins, and that a bad team can be a good one if it would just gain consistency (meaning play regularly the way they do when they win).

At any rate, losses are valid results, and so it's worth looking at losing streaks with as much attention as winning.

So far this season, the Wizards have losing streaks of 12, 8 and 5 games. The odds of a .200 team losing 12 in a row during an 82-game season: 99%. For a .300 team: 63%.

Over a 32-game schedule, a .200 team has a 78% chance of losing 12 in a row; a .300 team -- 25%.

If the Wizards manage to rip off several more wins in a row, it might be worth revisiting this topic. But, the winning and losing streaks the team has experienced so far are pretty normal for a bad team. 

Odds are that the Wizards will finish the season with a winning percentage of .200 to .300 -- 16-25 wins. I think getting to .300 (24-25 wins) is a reasonable goal for the team. They'd have to go 18-29 the rest of the way -- a .383 winning percentage.

That would be progress in a season that's already lost.