Daily Fantasy Hockey Strategy: Why You Shouldn’t Care About Home/Road Splits
With only a three game NHL slate tonight, I won’t be playing DFS or doing my normal daily picks write up. But I can’t stop thinking about hockey. I love hockey. So today I’m turning my attention to a more general issue of daily fantasy hockey strategy: home/road splits for individual players.
A Twitter friend piqued my interest last night when he went with a stack of Edmonton’s top line on the road in Minnesota. He cited the line’s elite performance on the road, and the line has indeed been much better away from Edmonton. At home the line averages a combined 4.2 points per 60 minutes, but on the road they average 7.2 points per 60. That stack delivered for my friend last night as they were responsible for both of Edmonton’s goals and totaled five points on the night. And that isn’t the only stack that has delivered for this particular friend recently. In the last week he’s had three big finishes in GPPs. For that reason, I’m hesitant to question anything he does.
That said, I’m very skeptical of reliance on home/road splits. As an example of my skepticism, Edmonton’s two-goal scorer from last night, Benoit Pouliot, is averaging one point per 60 minutes at home and 2.8 points per 60 on the road. Why? Because his shooting percentage on the road (22.9%) is more than double what it is at home (11.1%). As the great Eric T. once said in regard to shooting percentage: whatever happened last month or even last year is pretty close to meaningless. As it says in that piece, even a full season’s shooting percentage data is “almost entirely noise.”
Even when Eric looked at individual shooting percentage over a three year period for players who played more than 2500 minutes, the correlation between their shooting percentage in that three-year sample and their shooting percentage in the next three years is not overly strong. Pouliot has played just 21 home games and 18 road games this year, so it’s clear that his shooting percentages in such small samples have absolutely no predictive value.
I took a look at the 543 skaters who have played at least 200 minutes both at home and on the road this year and pulled out the 77 players whose difference in home/road points per 60 is one or more standard deviations from the mean. Of those 77 players, 64 have at least a three percentage point difference between their home and road shooting percentages. The average difference in absolute shooting percentage for the group is 7.31%. Counting on those differences in shooting percentage to keep up seems tenuous at best.
If I were trying to come up with a reason why a player might have a better shooting percentage at home or on the road, I might posit that a player could get more favorable or less favorable matchups at home or on the road. As an example, a home coach with last change could have a tendency to use his top line against the opposition’s top line giving them a tough matchup, or he could have a tendency to put his top line on the ice against a depth line thus giving them a more favorable matchup. On the road, the same idea is true, but it seems unlikely that a player is going to see significantly more favorable matchups than bad or vice-versa. Given that those matchups are largely out his team’s control, you have to assume a player is going to see an average number of tough and favorable line matchups on the road.
However, the individuals on the ice opposite a given player likely have very little to do with his shooting percentage anyway. It seems counter intuitive to say that the quality of the opponents on the ice has little to do with how likely a shot is to find the back of the net, but that’s reality. As you can see here and here, defenders have very little control over how often shots against turn into goals. They have much more control over how many shots are attempted when they are on the ice, but they have little control over what happens to the puck once it leaves a shooter’s stick. That’s not to say they have no control, but the level of control they have is not significant enough for us to have any confidence in that level of control continuing in the future. As a result, it wouldn’t even matter if a player actually did see more “favorable” matchups either at home or on the road.
None of this means that Pouliot isn’t going to score again in Edmonton’s next road game. Or even that Pouliot is going to trend hot at home and cold on the road from here on out. It just means that over time, a long period of time, these things will even out. The rest of the season isn’t nearly long enough for that to happen. I’m not saying you should bet on Pouliot at home going forward, I’m just saying you also shouldn’t bet on him just because he’s playing on the road.
I linked to four pretty geeky, stat-heavy articles in this post, and I wouldn’t blame you at all if you don’t want to take the time to read and attempt to understand them. Just know that the guys who wrote those pieces are smarter than you. They’re smarter than me. That’s why I take seriously what they’re saying. And that’s why you shouldn’t care about home/road splits.