Best you can do is go through the game to game logs, or find someone that already has (Try at Yahoo's HAG_list group or somewhere like that). At least you would only need to do it for one team per season for most of them.
Best you can do is go through the game to game logs, or find someone that already has (Try at Yahoo's HAG_list group or somewhere like that).
Thank you. I actually some half year ago became a member there, but didn't have Excel, and then I forgot about the site. Now I have Excel and view the content more easily. I didn't find what I'm specifically looking for here, but I did find some other interesting data, like save percentage for older seasons (I know it may not be totally reliable, but still).
I use gamelogs for some purposes, but not yet for goalie stats. Maybe I'll download all the available ones at nhl.com and insert its data into my tables, but I know from experience that even nhl.com's gamelogs are flawed (games are missing, game data is on rare occasions inaccurate). So if possible, when working on season stint totals, I prefer official summaries.
At least you would only need to do it for one team per season for most of them.
If I understand you right, it's actually more than that. It's on average 6 teams per season that a goalie that played for multiple teams during the season. There are 79 cases in 13 seasons.
From nhl.com, SHSA (short handed shots at) are available, as are TotalSA. So if I for those teams mentioned above, take TotalSA minus SHSA, I have the ESPPSA. Since all kinds of situational GA are available, I can calculate save percentages for SH and ESPP. However, despite league average save percentages when playing ES or PP are relatively similar, there will be some differences between factual ES save percentages and the combined ESPP ones. (I know that by studying the other about 24 teams per season.) Still, the differences are small, and the stats probably are "biased" in the first place by things like luck and circumstances. Maybe the combined ESPP stat is about as accurate for ES as the ES stat itself is.