Differences between hockey-reference/corsica/NaturalStatTrick

[GellMann]

Hey guys, how come these different sites have different numbers sometimes? For example, I was looking at Jack Eichel's pages, and they all say quite different things for Corsi, from 46% to 50%. I was under the assumption that none of them score-adjust (Corsica does if you ask it to) so these should all be the same CF/CA numbers, but the CA numbers on these sites differ by as much as 10 through 5 games. What gives, which one is right? I assume this happens for every player then since there's nothing special about the random one I picked out.

Click to read more...

 

[SniperHF]

Hey guys, how come these different sites have different numbers sometimes? For example, I was looking at Jack Eichel's pages, and they all say quite different things for Corsi, from 46% to 50%. I was under the assumption that none of them score-adjust (Corsica does if you ask it to) so these should all be the same CF/CA numbers, but the CA numbers on these sites differ by as much as 10 through 5 games. What gives, which one is right? I assume this happens for every player then since there's nothing special about the random one I picked out.

Just a quick look At NST and Corsica, Corsica has Eichel at 50% CF% but only through 4 games. They haven't updated him for his 5th game.
NST has 5 games.

Dunno what's up with Hockey Reference.

 

[GellMann]

Just a quick look At NST and Corsica, Corsica has Eichel at 50% CF% but only through 4 games. They haven't updated him for his 5th game.
NST has 5 games.

Dunno what's up with Hockey Reference.

Didn't see that NST updated the fifth game, but when it was at 4 I don't believe it was in agreement with Corsica, it was off by a shot or two.

But yeah, HR is the big one, because I'm now remembering seeing that site be wrong about a lot of things, including PP%, goals for/against for teams, etc.

 

[SniperHF]

But yeah, HR is the big one, because I'm now remembering seeing that site be wrong about a lot of things, including PP%, goals for/against for teams, etc.

FWIW NHL.com agrees with NST.
68 and 72.

 

[GellMann]

FWIW NHL.com agrees with NST.
68 and 72.

So the lesson is probably to be skeptical of hockey reference.

Edit: Yeah, looking on the Sabres' franchise page, they even give two different values for team goals/team goals against in different places on the same page

Shame because i love that site

 

[Mickey Marner]

Hockey-reference uses even strengh minutes which includes 6v5, 4v4, 3v3 etc. The other two are pre-set to 5v5 only.

 

[morehockeystats]

I don't know about NST and HR, but I remember encountering wrong data on Corsica.