Nice work! Thanks!
A month ago I posted in the BU thread that I was looking into quality starts vs really bad starts for BU goalies. To do that I needed all goalie stats for each NCAA season to determine the mean save percentage for each season. collegehockeystats.com had that information from the 2003-04 season forward, allowing me to find the actual mean save percentage for those seasons. For 1999-00 through 2002-03 I was able to come up with solid numbers, although they are missing some (2001-02 & 2003) or all (1999-2000 & 2000-01) CCHA goalie data. From 1973-74 through 1984-85 I used tops goalie (by GAA) stats, so well solid, the actual mean save percentage is certainly lower than the number I have used.
However, I was unable to come up with many goalie stats for 1985-86 & 1986-87 and 1988-89 through 1998-99, so I posted a request for help. As no one sent me any stats, I ended up extrapolating goals and saves for 1992-93, 1993-94 & 1994-95 and for the other seasons I used what limited sources I have or was able to find to create rough mean save percentages. After comparing them to other metrics I am confident that they are relatively accurate numbers for each season, but I would still appreciate goal and save information for CCHA goalies before 2002 and for almost all goalies before 1999.
Looking at the actual and estimated mean save percentages since the 1973-74 season shows the overall rise from around 89% for the top goalies to a high of just under 91.4% for all goalies in the 2014-15 season. However, last season it dipped slightly and so far this season it is down again, to just under 90.5%. That is down almost 0.9% from 2014-15 and is the lowest itís been since 2006-07.
After compiling BU goalie stats from 1973-74 (partial through 1983-84) forward I got sidetracked by the idea of quality starts so I also compiled game-by-game goalie stats for every game this season. I then looked at every goalie's number of starts, quality starts, really bad starts, and W-L-T for quality starts for the season through 12/18/2016.
In looking at all goalies by class freshmen are leading the way this season in starts and quality start percentage, with juniors just behind in both categories. Sophomores are third in starts, but last in quality start percentage, while seniors are last in starts, but ahead of sophomores in quality start percentage. In really bad starts juniors lead with just 12.8%, freshmen trail with 18.3%, sophomores have 22.0% and seniors 22.4%.
Freshmen also have the best mean save percentage and best starting goalie save percentage, both at 0.912. Juniors are again just behind them at 0.910 overall and 0.911 for starting goalies. Seniors are at 0.901 for both, while Sophomores come in last at 0.894 for both.
However, for the most part, freshmen have not been handed the starting job. Nine goalies have started every game, none a freshman. In fact the top 18 goalies in starting percentage include no freshman (4 seniors, 11 juniors, 3 sophomores), with BUís Jack Oettinger and Miamiís Ryan Larkin both tied at 19th with 2 other goalies. However, 5 of the next 6 goalies in starting percentage are freshmen (all above 80%) and another 7 freshman goalies have started between 60% and 80% of their teamís games.
I have uploaded a NCAA Goalies spreadsheet if anyone is interested in looking at it. The first sheet has a list of definitions and abbreviations used on the other sheets. The second sheet has every game appearance for every goalie (or almost every appearance Ė I may have missed a few relief appearances). The third sheet has every goalie's number of starts, starting percentage, quality starts, really bad starts and quality starts winning percentage. The fourth sheet has the mean save percentage for every season starting with 1973-74 and what each is based on. The fifth sheet shows game save percentage vs winning percentage for every decision earned by a starting goalie.
Nice work! Thanks!
Very interesting stuff! I wanted to come up with a new goalie stat in which i term GAF or Goalie At Fault. This stat would take into consideration goals that no matter what the goalie could have done it would have been a goal anyways (such as a deflection that went opposite direction into top corner etc). Basically, it would strip out goals that goalies had no chance on to give a better idea of GAA and save percentage. Of course it would be subjective just as a hit or error are in baseball but it would give some clearer stats for goalies in my opinion as they aren't punished for goals they had no chance on.
I have to look, but a guess would be that Kevin McCabe (Brown) got the shaft in the 1976 AA vote.
Really interesting info.
Just shows how much a solid goalie has to do with a teams actual success. Redmond for Tech has the highest QS % of goalies that have played a fair amount. Look at the difference in Tech's record before he played and after he took over. He's the single biggest reason they turned their season around. Bitzer the same for BSU, giving them a QS almost every night.
No accident those two teams are at the top of the WCHA when you look at the goaltending they have gotten.
I finished updating my 2016-17 NCAA Goalies spreadsheet with full season stats for all goalies for 2016-17. I have also made some changes to the spreadsheet, adding Quality Start vs Quality Start (when both starting goalies had a quality start) and Other (all games not a quality or really bad start). I also added W-L-T records for each category to give an idea how successful a goalie was when having a quality start, quality start vs quality start, really bad start and other start.
Since quality starts are determined by the mean save percentage for a season I will start with it. When I uploaded the first version of the spreadsheet in December this seasonís mean save percentage was about 90.5%, while the final mean save percentage was 90.822%. Still, thatís the lowest itís been since the 209-10 season (and just a tad higher than 2010-11ís 90.818%), 0.4% lower than 2015-16ís 91.22% and 0.55% lower than 2014-15ís all-time high mean save percentage of 91.37%. The decrease in save percentage correlates to a rise in goal scoring, up 506 goals from 2014-15 and 213 goals from 2015-16.
The mean quality start percentage was 55.3% for all goalies and 57.5% for goalies identified as their teamís starting goalie(s). I identified 48 teams as having one starting goalie, 11 having a 2-goalie rotation and one team (Michigan) using a 3-goalie rotation. For teamsí starting goalies the percentage ranges from a high of 82.9% to a low of 26.1%. The higher the percentage the better, but you also want to see a really low really low Really Bad Starts percentage as well. And that is what we see with the top 2 starting goalies in quality starts, 2 of the 3 lowest really bad starts percentages. That makes them the most consistent starting goalies for the season.
I added W-L-T because I was interested in seeing how well a goalie did when having a quality start vs other goalies that also had a quality start. Overall Quality Starts vs Quality Starts should total to have equal wins and losses since when one goalie wins one goalie loses. In 2016-17 there are 3 more wins than losses because 2 loses by SNHU goalies are not included (they are DII) and Collin Delia had a no decision in a game he played just under 50 minutes in that I gave him a quality start and his relief got the loss (see note 1).
Looking at quality vs quality starts shows what goalies had the most, which makes it harder for them to have as good a record as goalies that didnít have to face as good opponent goaltending. The goalies with the most quality vs quality starts were Providenceís Hayden Hawkey and Bemidji Michael Bitzer, each with 18. Armyís Parker Gahagen was 3rd with 17 and Denverís Tanner Jaillet was 4th with 16. The goalie with the most wins was Bitzer, with 11, followed by Air Forceís Shane Starrett, Penn Stateís Peyton Jones and Hawkey with 9 each. Starrett was almost perfect when having a quality start vs a quality start, going 9-0-1, with Jones 2nd, going 9-1-0.
When you widen out to also look at quality starts you see that Bitzer had 28 quality starts and went 21-4-3 (0.804) in those games, with 18 of them against quality starts by the other goalies in which he went 11-4-3 (0.694). Jaillet had 31 quality starts and went 22-5-4 (0.774), with 16 of them QvQ starts in which he went 7-5-4 (0.563). Gahagen had 29 quality starts and went 16-10-3 (0.603), with 17 QvQ starts were he was 4-10-3 (0.324). Starrett was almost perfect in quality starts, going 21-0-3 (0.938) in 24 games, but only had 10 QvQ starts, in which he was the aforesaid 9-0-1 (0.950). In fact , 230 of the 241 loses (95.4%) by starting goalies having quality starts were QvQ games, so a goalie that was able to win most of them did very well.
I also showed how often a goalie with a quality start faced another goalie that had a quality start. For starting goalies it ranges from Colorado Collegeís Alex Leclercís 92.9% (13 of 14 quality starts) to UMass Lowellís Tyler Wallís 20.8% (5 of 24 quality starts). For the top 19 goalies with 20 or more quality starts it ranges from Providenceís Hawkeyís 75.0% (18 of 24) to Wallís 20.8%.
Looking at quality starts and quality start percentage, freshman lead in starts, but were edged out in percentage by juniors, with sophomore and seniors remaining in the same places. Looking at the mean save percentages, from December to the end of the season freshman dropped off slightly, from 0.912 to 0.909, while juniors improved slightly from 0.910/0/911 to 0.911/0.912. However seniors improved from 0.901 to 0.909 and sophomores improved, from 0.894 to 0.901/0.902.
Five goalies ended up starting every game for their team, down from nine in mid-December. The leading freshman was Hunter Miska, who was 15th in games started percentage, one of 6 freshmen in the top 25. Of the 48 teams using primarily a single goalie 15 were freshmen, 10 sophomores, 15 juniors and 8 seniors.
Overall, of the 989 games won by a starting goalie 783 (79.2%) of them were won by a goalie who had a quality start, so the importance of having a goalie who can have consistent quality starts is extremely important.
There are currently 1 users browsing this thread. (0 members and 1 guests)