A famous-but-unknown sports blunder

19 Jan, 2026

One of the most well-known and oldest strategic plays in sports is pulling the goalie in hockey to get an extra skater on the ice when you are losing in the waning minutes of a game.

There is wide debate over when coaches should pull the goalie—many stats nerds think they wait too long—but no one disputes the underlying logic of the strategy. In normal 5 on 5, teams have about 0.65% chance of scoring in any 10-second period. With the goalie pulled you increase your chance of scoring to about 1.97% in any 10-second period and you increase the chance of the other team scoring to 4.3% in any 10-second period.

This is, of course, a -EV strategy in terms of net goals. But it’s +EV play in terms of winning the game. If you are losing by 1 late in the 3rd period, at some point in time you are happy to trade an increased chance of losing by 2 for an increased chance of tying the game. A loss is a loss, whether by 1 or 2. But tying the game is highly valuable relative to losing by 1. So you are ready to take risks.¹

Essentially, at some point a goalie becomes negative EV relative to a 6th skater for the purpose of winning the game.

And the EV of such a move increases as time remaining heads toward zero. In the limit—say you have a faceoff in the offensive zone with 1 second remaining—it becomes literally impossible for the other team to score, since they couldn’t get the puck down the ice in time. At that point, it becomes +EV to put a 6th skater on the ice in a tie game. ²

This isn’t rocket science. And it didn’t take the Moneyball revolution to figure it out. They started doing it in the 1930’s. And it happens at the end of literally every one-goal (and most two-goal) NHL and college hockey game.

Which made it all the more surprising when I was watching a replay of the 1980 Miracle on Ice US-USSR Olympic game, and I realized that the Russians didn’t pull their goalie in the waning minutes of the game. Down 4-3, they simply left him in. Why? Sergei Starikov, a player on the 1980 Soviet team, has offered a simple explanation. "We never did six-on-five," not even in practice, because "Tikhonov just didn't believe in it."

Victor Tikhonov was, of course, the Soviet coach, a famous disciplinarian but also the inheritor of the ingenious hockey innovations of Anatoly Tarasov, who invented the modern style of play and had Russian national teams using it almost a generation before western teams, one of the key reasons Russian teams dominated international play in the 60s and 70s.

This is a mistake of epic proportion. And probably cost the Russians somewhere between 5-10% chance to tie the game. I cannot even imagine the internet uproar that would occur today if a team did this next month in the medal round.

Tikhonov is well-known for one goalie-pull in the Miracle game: yanking the best goalie in the world, Vladislav Tretiak, at the first intermission, and putting in his second-stringer, after Tretiak gave up a bad rebound in the last seconds of the first period, allowing the US to tie the game at 2. He subsequently has called it the worst decision of his coaching career. Hard to believe not pulling the goalie in the last minute of the game could be far behind.

A second question is how have I never noticed this? I’ve watched the Miracle game at least 10 times in full, and I’ve seen the final 2 minutes countless times. One answer is that the camera shots are very tight, and the puck never leaves the US half of the ice in the final minute. We never see the Soviet goalie in the frantic final moments. The other is that color commentator Ken Dryden does mention prior to a faceoff with about 1:15 to go that the Soviets must be thinking about pulling the goalie soon, but neither he nor Al Michaels make another mention of it as the clock ticks down.³

——————-

#EV #hockey