Backups, Babcock and SEGABABAs

You can never really know why something a hockey coach does starts to annoy fans sometimes. Well, you can usually guess. Look at the results. Were they good or bad, and if they were bad, you’ll find a lot of people deciding independently that whatever the coach did was obviously wrong and should be changed. It’s not exactly unusual to find coaches who make decisions the same way.

Player on a shooting draught? Fourth line! Goalie with some bad results? Waive him! Turn the puck over? Benched. Take two penalties in one game? Press box for life! Most people can come up with examples of extremely results-based coaches doing things that had no relationship to the reality of a player’s actual play. Mike Babcock was not like that.

He was so much not like that, that fans got angry over the things he wouldn’t just arbitrarily change. This year’s massive annoyance became the way Babcock utilized the backup goalie in back-to-back games. The Leafs had a lot of them front-loaded into the schedule, so there were ample anecdotal results that could be added up into anecdata as proof that Mike Babcock was wrong about this, and finally, finally, oh glorious day, Sheldon Keefe is going to fix it, just like he made the point shots go in.

So, let’s backup the backup truck, and look at what Babcock was really doing, and see how this is different.

Mike Babcock made a blanket policy that the starter played the first game of the back to back, and the backup always played the dreaded SEGABABA (SEcond GAme of a BAck to BAck). This was carried out regardless of the other circumstances: team strength, travel, home or road splits or anything else you might think of.

The general idea behind making this decision is that for the Leafs, with a starter who is well above average and a backup (and this applies to the entire list of them over the last two years) who is performing well below average, the way to maximize the points out of these two games is to play the starter in the first one, where the chances of winning it aren’t impacted by the real effects on performance of a SEGABABA.

For a replacement level goalie, remember, he is performing below league average overall. His chances of winning even the softest of games is very poor. His main function, and the thing that adds value to the team, is to give the starter a day off. Any arbitrary rule about which order to play them in on back-to-backs is likely as good as any other.

This season, so far, the second half of the back-to-back has always been a road game with the first one a home game. All but the very first pair, that is, which came in the first week of the season. That first SEGABABA was at home to Montréal, and the Leafs lost in a shootout after winning the road game the night before. That was the last time the back-to-backs yielded any points out of the second game.

By making it a blanket rule, where the only consideration was playing the starter behind the fresh team to try to get the two easier points that way, Babcock picked a simple method that wasn’t open to any sort of jockeying around that, and I’m speculating here, might make the various goalies resentful or indicate to the skaters that they weren’t expected to win in a particular game because suddenly the backup was in. Essentially, the onus was on the skaters to perform well enough while tired to help the poorer goalie get something out of the SEGABABA.

This season, the Leafs failed spectacularly to do that, producing efforts so bad that Michael Hutchinson and Kasimir Kaskisuo have the second and third worst Expected Save % (all-situations via Moneypuck) in the NHL so far. Expected Save % is what a league average would produce on shots of the number and quality faced. Only the poor, beleaguered Laurent Brossoit in Winnipeg has had a tougher job. Faced with that mess in front of them, even if the backups had not performed extremely poorly, the Leafs still would have lost those games. By playing so poorly, all the backups did was make a bad situation into almost the worst outcome you can have in the six SEGABABAs they’ve played.

If you really wanted to maximize the probability of getting four points out of these two games, you would need to consider the opponent’s travel and rest, the team strength, and a host of other factors. Given the variance in even a good backup’s results, that all seems like time better spent on other concerns. Just cut it down to a simple set of guesses about what will work best.

The more complex rule than Babcock’s, and the one I’ve always favoured, is to add in the home and road effects on win probability and make the choice that way. Home ice advantage adds a similar amount to a team’s chances of winning as a SEGABABA takes away. There is a lot of parity in the NHL this year, and unless you’re playing the Red Wings, you are going to get some kind of fight out of any team, so any team can beat you on any given night. Obsessing over wins vs the so-called division rival is a lot harder to do when there are five of them, so that’s also layering on too much clever scheming for my taste. You need the most points as the first priority, keeping them away from all your division mates is secondary.

My rule considers two factors over and above the base probability that the team matchup has. So if the first game is at home and the second on the road, the SEGABABA effect and the road game effect double up and make that second game extra-hard to win. You’re left with really only one rational option in my opinion, and that’s to play the starter in the first game, maximize your home ice advantage and let fate take care of the second game outcome. This is the Babcock method, and I like it. And no one should be expecting the backup to “steal” any wins. They shouldn’t ever, given the quality of the goalie performances right now.

When the home and road split is reversed, however, I would prefer to mix this up. Play the backup on the road in the first game, and the starter at home on the second. Now the two effects are cancelling each other out in both games, or near enough. Each goalie gets a fair chance at a win, with the starter far and away more likely to actually get it, and that is playing the probabilities correctly to get the most points you can.

It’s not about changing what Babcock did because it didn’t work x number of times, or feeling sorry for the poor, sad backup who had a tough outing, it’s about some simple gambling strategy, and the understanding that even if you’re making the smart bet, you might lose. Losing isn’t a reason to change just for the sake of it. This in’t the roulette table, and you don’t need your lucky socks to win.

By the way, so far this season on the Marlies, Joseph Woll (the backup at this point in his rookie season) has played the SEGABABA every time, but most of the back-to-backs in the AHL are in the same place against the same team.

With this decision to play Hutchinson on the road in Buffalo and Frederik Andersen at home the next day, it appears that Keefe is using a modified approach to deciding goalie starts, but it’s still based in the same principle that led to all that consternation over Babcock and his unwillingness to change for the sake of change.

Keefe gave these reasons, all but the last of which make a lot of sense.

Now, if the Leafs just had a backup they were a little less afraid to use at all, this would be a lot less contentious an issue.