View Single Post
Old
12-07-2012, 05:48 PM
  #71
Czech Your Math
Registered User
 
Czech Your Math's Avatar
 
Join Date: Jan 2006
Location: bohemia
Country: Czech_ Republic
Posts: 3,650
vCash: 500
Quote:
Originally Posted by Rowdy Roddy Peeper View Post
That's my point. The ESGF/GA and On/Off ratios vary so wildly relative to what most every reasonable hockey observer would consider better seasons (check out Elias '01 or LeClair '97 relative to the best of Jagr and Lemieux) that it's hard not to question their validity in this debate.
.

Your point is a good one, as those ratios do vary wildly. I think that's because there's much more luck (factors outside the player's control) than there is in points/PPG. The player has much more direct influence over points/PPG than he does over those ratios. They do give us some important additional info, but the context must be considered and they are most useful over multiple seasons.

Quote:
Originally Posted by Rowdy Roddy Peeper View Post
How would one then rationalize the ESGF/GA ratios of Lemieux and Jagr in their best seasons relative to the all the other players who had played full seasons and had much better results?
As stated, the ratios are much less in the player's control than points/PPG and there is much more luck involved. The quality of linemates is always part of the context and, in the case of superstars, may influence the ratios much more than those superstars' points/PPG. The quality of the defense and goaltending directly influence the ESGF/GA ratios to a large degree, while the quality of other lines on the team directly influence the On/Off ratios to a large degree. On a single season basis, they are particularly useful for demonstrating that a player did not have an effective possession game and/or abandoned defense to a large degree (see Gretzky '94 or Lemieux '03). Over multiple seasons the ES data has much more validity IMO.

Quote:
Originally Posted by Rowdy Roddy Peeper View Post
That makes sense. However, it should cut both ways. If you believe that a very productive player playing on a different line would likely be a considerable disadvantage to one's On/Off ratio, doesn't it follow that a very productive player on the same line would likely be a considerable advantage to one's production?
There's context to consider for each category of data. Again, the context tends to matter less for superstars' points/PPG than it does for the ES ratio, especially for single seasons or less.

Quote:
Originally Posted by Rowdy Roddy Peeper View Post
I do think it's the best single indicator for all the reasons you mention. After seeing the results, it's hard to give much credence to the ESGF/GA data.
I understand your position and agree that ES data has much less validity on a single season basis. However, all data requires context. Points/PPG may be substantially affected by the player's role on the power play and the quality of teammates on that PP. That doesn't invalidate points/PPG, but it's something to consider

Quote:
Originally Posted by Rowdy Roddy Peeper View Post
I completely disagree. Beginning at the player's first game of the season is the least arbitrary point possible - everyone has the exact same starting line.
That's true, but remember that players on generally pacing themselves for a full season. We do not know how well Crosby could have or would have maintained his production over a full season, if he had remained mostly/fully healthy. What we do know is that all players also have the exact same finish line, and Crosby dropped out halfway through the race. Anything less than a full season is somewhat artificial, in that it's attributing the characteristics of a full season's production to production in a shorter time frame, when other factors (balanced schedule, luck, etc.) are less likely to even out and present a level playing field.

Quote:
Originally Posted by Rowdy Roddy Peeper View Post
I'm not sure who based an argument for Crosby on ES production, but it wasn't me. It seems reasonable that superstars with superstar linemates would fare better at ES than superstars with blunt instrument linemates.
I introduced the ES data, in response to claims that Crosby's "overall game" over a half season was particularly unique, even more so than his offensive production. I believe it's probably the best data to at least attempt to measure some degree of the player's "overall game." The data is much less reliable for a single season, let alone a half season, but it suggests that Crosby's "overall game" for that half season was nor more unique than his offensive prodution, perhaps even less so (given that other players' ES ratios were based on full seasons when their PPG were lower than in half season).

Czech Your Math is offline   Reply With Quote