The impact of rest on recent NHL shot rates

hockey

#1

I’ve been poking around with trying to quantify just how much rest affects the way NHL games are played.

In what follows, “0 days of rest” means “played the previous night”, “1 day of rest” means “played the night before that” and “2 days of rest” means the most recent game was at least two nights ago. Red is more shots than average, blue is less.

This is the shots the home team gets, relative to what you would expect given the team shot rates of the teams playing in each game in the last five NHL regular seasons. Mostly you see red here - home teams get more of the shots. Unsurprisingly, well rested home teams generate a pile of offence (lower-left) against tired road teams. More surprisingly, it seems as though a two-day differential (at least) of rest is required to make this kind of difference.

Similarly, at least a two day differential in rest is needed to make a serious dent in the offence the home team generates, and it doesn’t seem to hurt them as much as it benefits them when the shoe is on the other foot.

The road team’s offence hardly seems affected by rest at all - when they have a hefty rest advantage on the home team, they do well, but other than that are almost completely uniform.


#2

If you mask out home-ice advantage, it’s much less colourful and much easier to read:


#3

You mentioned on Twitter that this is only reflected in 10% of season games, which we can see are the ones in the corners, so thanks for that!

I’m also curious what variation from average do we see in those specific games? +5%, +10%, +20%, etc greater than an equal rest game?


#4

The top right images (heavy rest advantage for the road team) is 4.4% fewer shots for the home team and 6.2% more shots for the road team. Roughly speaking, the roles of home and road teams are reversed. So, you shouldn’t think of games like that as schedule losses, you should just think of them as road games.

The bottom left images (heavy rest advantage for the home team) are not as impressive; 3.8% more shots for the home team, 3.2% fewer for the road team. It’s almost but not quite as much as getting home ice advantage “again” as it were.


#5

How many extra shots per game does that work out to in each situation? 1-2?

It’s really interesting to me that the major advantage (or disadvantage) seems to come on the most dangerous types of shots. I would have guessed that even though you might get more they’d be more evenly spread out over the rink, but this seems to suggest that rest gives you maybe one extra “good” chance per game.


#6

Average unblocked shot per hour at 5v5 is around 40 so yeah, about a shot or two, sometimes three. Shooting percentage on shots like that from the slot is in 15-20%, though, so it’s a pretty hefty advantage still.


#7

That’s actually a good metric to add to the chart, imo: the overall shot impact for games in those situations, i.e. +2 shots or +0.3 xG, etc.


#8

Neat. I like this as a kinda piece to the “starting goalies B2B” discussion and would be interested to see goalie performance in these games (ie tired goalie/rested goalie). If shot rates in most cases arent dramatically different - should a coach even take this into account when deciding on which goalie to play? If a home team is likely to get the shot share against all but a really rested away team - does it make sense to start your “rested” goalie (who may be your “worse” goalie) and give away some of that natural shot advantage? Or do you take the chance on the “tired” goalie, who may see a drop in his expected save% percentage blahblahblah.

I guess this suggests to me “rested vs tired” team shouldnt be a big factor in a coach’s goalie decision.


#9

I hadn’t thought of it from a goalie-centric point of view but I don’t think I adjust my goalies too much based on a handful of high-danger shots at most. I think I’d rather have a baseball style “rotation” where everybody has their nights ahead of time and they know their routine and can prep for it.