clock menu more-arrow no yes mobile

Filed under:

Approximating Touches for the Oilers This Season

What if we used the real-time stats for good instead of evil?

By ryan_batty
Derek Leung

As a rule I'm not a big fan of the real-time stats tracked by the NHL. Hits, blocked shots, giveaways, and takeaways can provide some interesting information but they also tend to be very misleading. I've heard many times that Jeff Petry is a terrible defender and his giveaway total, third highest among NHL defencemen, is evidence of that. It's true that Petry has a lot of giveaways, but look at the top ten, how many would you want on your team? Giveaways obviously aren't good but maybe that number really just tells us that the more a player handle the puck the more likely they are to turn it over from time to time. The same goes for hits and blocked shots, both are nice but they also mean that you don't have the puck, and as long as the goal of the game is to score more than your opponent then I tend to think that having the puck is important.

So more often than not I simply ignore these numbers. But last night I came across this conversation on Twitter:

An interesting idea for sure. Take those real-time stats, look at them in a slightly different light and maybe we can use them to approximate how many times each player touches the puck. I think it's fairly obvious that this method misses a lot of touches but it should provide a rough approximation of which players are handling the puck the most. The equation for touches would therefore be:

Touches = assists+shots+missed shots+blocked shots+hits against+giveaways+takeaways

So, if measured this way, how would the Oilers look this season? (To save space I combined the Corsi events into a single column)

FORWARDS
Player A CF HitA GvA TkA Touches /GP /60
Hall 51 424 159 100 76 810 10.80 33.23
Eberle 36 354 130 71 61 652 8.15 25.60
Jones 4 106 71 18 15 214 4.12 25.46
Perron 29 377 114 42 52 614 7.87 25.26
Joensuu 2 75 49 18 14 158 3.76 24.53
Pitlick 0 18 14 1 2 35 3.18 23.57
Hemsky 17 162 80 37 33 329 5.98 22.57
Yakupov 13 217 56 32 18 336 5.33 22.54
Gagner 26 250 70 47 28 421 6.28 20.95
Arcobello 14 127 38 13 18 210 5.12 20.63
Acton 2 36 22 4 15 79 2.63 20.26
Nugent-Hopkins 36 298 77 50 63 524 6.55 19.75
Hendricks 0 79 51 9 9 148 4.48 19.64
Smyth 13 199 79 35 27 353 4.90 19.61
Gazdic 2 51 41 11 6 111 1.66 17.44
Gordon 13 140 97 19 21 290 3.92 16.44
Lander 1 39 29 9 6 84 3.11 14.70
Brown 0 5 3 0 0 8 1.00 10.53

DEFENCE
Player A CF HitA GvA TkA Touches /GP /60
Larsen 8 98 29 15 4 154 5.13 18.44
Petry 10 217 112 96 29 464 5.80 16.57
J. Schultz 21 237 106 63 20 447 6.04 15.97
Belov 6 132 51 46 7 242 4.25 15.63
Ference 15 148 135 55 13 366 5.15 14.87
Klefbom 2 31 12 9 3 57 3.35 13.87
Potter 5 23 12 6 4 50 3.13 13.59
N. Schultz 4 60 107 35 2 208 3.47 12.38
Marincin 6 75 25 33 17 156 3.55 11.62
Fraser 0 21 23 17 4 65 2.83 11.54
Smid 1 16 17 10 5 49 2.88 9.77

A couple of thoughts:

  • It really doesn't matter what you look at, Taylor Hall is a beast. This team will go as far as he takes them.
  • Centres appear to be under counted by this method. If I had to guess why that is, I'd say that it's probably a result of their roles in the defensive end of the rink.
  • Philip Larsen looks just fine by this measure, but has the third worst CF% among this group, so like all stats, context matters.
Like I said above, this method misses a lot, but it's interesting nonetheless.