View Single Post
Old
12-07-2012, 02:09 AM
  #69
Czech Your Math
Registered User
 
Czech Your Math's Avatar
 
Join Date: Jan 2006
Location: bohemia
Country: Czech_ Republic
Posts: 3,490
vCash: 50
Quote:
Originally Posted by Rowdy Roddy Peeper View Post
Thanks for the explanation, and sorry for the lateness of my reply here.

Correct me if I'm misinterpreting the numbers, but wouldn't they suggest that Crosby's '11 ESGF/GA ratio and On/Off ratio were better than several of the names at the top of that list, including Lemieux and Jagr in '96?

If that's true, it wouldn't necessarily be the most accurate indicator of the best seasons, would it?
I'm uncertain which list you mean, but Crosby's ESGF/GA ratio and On/Off ratio were better than some players with higher PPG, just as some players with lower PPG than Crosby had better ES data. There's a couple of things to keep in mind:

1. The ESGF/GA and On/Off ratios are only available for a full season. Since most players have a substantially higher PPG during their best 41+ games than they do over the same full season, it would stand to reason they would have substantially higher ES ratios during their "half seasons" than over the same full seasons.

2. In the case of Lemieux & Jagr in '96, the "Off" part of their On/Off ratios were comprised in large part by the other's line. Suffice it to say that a prime Jagr or Lemieux is a much tougher comparison than an injured Malkin having an off year.

Is (adjusted) PPG the best indicator of which player had the best "half season"? Certainly other factors can and should be considered (such as the ES data), but if one was allowed only one stat with which to gauge the quality of a forward's half season, PPG seems the best one. It's the most direct measure of the player's offensive contribution, as the others are influenced by additional factors (quality of defense/goaltending, quality of other lines) which are outside the player's control.

Quote:
Originally Posted by Rowdy Roddy Peeper View Post
It seems cherry-picked to me. We aren't picking the best "__" game segment to choose from Crosby, we're using all of the games he played that season. Using a streak of all the games a particular star played from the start of the season would be the best way to put everyone on a level playing field, I would think.
It's a matter of interpretation. Some, such as yourself, may believe that using any 41+ game stretch in comparison with the first 41 games is cherry-picking. Others, such as myself, would say that 41+ consecutive games is basically apples to apples, especially as the games seem to be as tough or tougher in the second half as the first half. Also, even restricting it to 41+ games, rather than a full season is somewhat artificial. Players are generally judged on at least a full season basis, as the schedule balances out over a full season and the larger sample means the "luck" factor is less of an influence (which is one reason why multiple seasons are even more reliable than a single season). At some point, restricting it not just to 41+ games, but only the first 41+ games of the season is reminiscent of "only N players have X goals, Y assists and a PPG of Z or better." It's not that it isn't impressive for the player to be part of such an exclusive list, but that the more the criteria are arbitrarily tailored to meet that player's stats, the less impressive it becomes.

Quote:
Originally Posted by Rowdy Roddy Peeper View Post
All great years. But it is instructive to note that Lindros and Selanne both had First-Team All-Stars on their line (and still scored at lesser paces), and Ovechkin's center was 4th in league scoring. Crosby's linemates were Dupuis and Kunitz, who wouldn't be mistaken for First Team All-Stars or Top 5 scorers by a blind man in a dark alley.

Scoring at the pace he did with what he had to work with was always a big part of what made it so impressive.
2010 Ovechkin scored at a higher rate (both raw and adjusted) to start the 2010 season and doing so for 52 games makes it all the more impressive. 1999 Selanne (on an adjusted basis) and 1995 Lindros scored at lower rates, but not that much lower. They did have better linemates than Crosby, but they also had better ES data that was at least similar and often much better during those years.

Czech Your Math is offline   Reply With Quote