Post-Game Talk: Game #25: Anaheim Ducks @ Colorado Avalanche , 6pm PT, BSSC

[duxfan1101]

IMO one of the major flaws with fancy/possession stats is that they only measure 5 on 5. Making value judgements while leaving out special teams is lame.

It's not that they don't measure special teams (anyone can go on Natural Stat Trick and find fancy stats for special teams within a couple of minutes), it's that you have to report those stats separately because those minutes are completely different than 5 on 5 minutes. Your xGF% on the powerplay will usually be in the 80-95% range, which makes sense because you generate way more offense than you give up, and the opposite applies to PK. So if you simply report a player's xGF% in all situations in a game, it will be significantly influenced by how much PP or PK time they get, and then the number really doesn't mean anything. Even strength numbers are also much easier to interpret. On average, you would expect a player to get around 50% of xG, so you obviously want to be above the 50% number. Nobody would be able to tell you (unless they have researched it) what a good percentage would be on the PP or PK.

 

[JAHV]

If you're looking at fancy stats for the PP or PK, that's where, as someone mentioned earlier, you want to look at rates or at production relative to teammates and opponents. If a guy produces 5.5 expected goals per 60 minutes on the power play, that's great, unless his teammates are producing an average of 7.5 expected goals per 60 on the PP. Or if the league average is 8.5 g/60.

As with all things, context is key. Advanced hockey stats do a better job than traditional stats, in my opinion, of putting a player's performance in the proper context so it can be evaluated against other players on a level playing field. But they still should not be used as blunt instruments.