When analyzing goaltenders uniquely GVT seems to paint an accurate picture when compared to canonically assumed knowledge. (It seems less accurate for forwards and unreliable for defensemen.)
The general gist of the statistic is goals prevented compared to a replacement goalie (slightly below average goalie). Related to save percentage, but there are some differences between the two statistics.
(I don't know what those random letter combinations mean either. GVT is the only stat I looked at. I do know SGVT is shootout GVT and is therefore useless and included in the final GVT.)
I don't think it's schedule adjusted unfortunately. Less important for regular season numbers, but significant for playoff numbers, so when analyzing playoff numbers I multiplied by a factor of:
1967-68 -> 1973-74 (4/3x)
1974-75 -> 1985-86 (27/24x)
Some sanity tests:
- Hasek is the greatest regular season goalie of all time. His playoff record while decent is comparable to quite a few other goalies (Parent, Brodeur, Belfour, etc.).
- Plante is arguably both the second best regular season and playoff goalie.
- Roy is the second best regular season goalie of his era, but there are a good number of comparable goalies historically. (Plante, Hall, T. Esposito). He's probably the greatest playoff goalie ever, where his average run during a 15 year period would be similar to Fuhr's cup winning numbers.
- Sawchuk has the greatest playoff run ever. A very good peak, but little else.
- Giguere's 2003 playoff run is one of the greatest in recent memory, and perhaps the best by a losing goaltender.
- Smith was an excellent playoff goalie.
- Osgood was an above average goaltender but not much else. He does have 2 very good playoff performances though.
- It seems to do a fairly accurate job of estimating 1st and 2nd All-Star Teams from the mid 50s to the mid 60s, except it suggests Bower got shortchanged quite a bit in this regard. (The top 100 list seems to agree with this assessment.)
Oddities (or maybe not):
- Durnan's numbers approach Hasek's, which would seem to be justified by his 6 1st team all-stars in 7 years. Maybe he should get more credit? Doesn't have the greatest playoff resume, but it's not that much worse than Hasek's. Extremely short career though.
- Hall doesn't seem to do as well as one would expect. He does consistently well throughout his career for the most part but never seems to as dominant as one would expect from a player with so many All-Star Team placements. (His numbers are still very good)
- Esposito fares exceedingly well during the regular season. Not sure if this should be the case, but if I had to guess poor competition might be the cause.
- Brodeur doesn't perform 'well' by this metric (goalies are being given some credit for influencing shot totals). He has some great seasons but suffers from inconsistency.
- Belfour comes out nicely in comparison a worse regular season goalie than Brodeur, but a better playoff one (arguably).
- Fuhr may be overrated in comparison to other dynasty goalies and some modern goaltenders. Quite arguably still a top 10 playoff goalie using this metric.
Goalies in top 100 list who have missing years:
- Tretiak (all)
- Gardiner (all)
- Brimsek (first half)
- Broda (first half)
Following placements were eyeballed. Prime (consistently great) and peak (consistently amazing for consecutive years) were weakly defined.
7 Best prime regular season goalies (~10 consecutive years):
7 Best prime playoff goalies* (~5 consecutive):
* Some will notice that these are mostly dynasty goalies. This shouldn't be surprising as they would have to be consistently great to win multiple Stanley Cups in a row.
** Belfour's excellent 5 year run was done on 2 separate teams, which makes it particularly impressive. (He did miss the playoffs once in-between though, but in his defense he was traded to the 2nd worst team in the league, however he did play poorly that year.)
*** Could very well be higher. His individual numbers most likely wouldn't be any higher than Belfour's, Smith's or Dryden's, but might make up for it in greater consistency over a longer period of time. He also lost some of his best years to the war.
7 Best peak playoff goalies (~2 consecutive years):
*Roy never really had great consecutive years but he had so many you can't really put him any lower.
By GVT, Hasek's regular season peak is comparable to any player in history, up there with Orr and Gretzky. That's a bit higher than Hasek is usually rated around here, but I'm not sure that it's wrong. It's why I had Hasek at #5 on the latest round of the HOH top-100 (although I didn't use GVT to get there.)
I would say Hasek has some separation over Durnan in GVT. While Durnan has the biggest single year of the two, it was a war year. And Hasek is well ahead in his 2nd best, 3rd best, etc seasons. Also, the thing about Durnan's numbers is that there are no shot counts available before 1952-53 (the current earliest season that the Hockey Summary Project). And I'm not sure if GVT uses the HSP data or not, so goalie GVT data may not include save percentage for some post-1953 years also. I don't know what assumptions GVT makes in separating team defence from goaltender performance in the pre-save percentage era, so Durnan may be getting some credit for team defence.
A note on Brodeur's playoffs - playoff GVT, unlike regular season GVT, does not include shot prevention credit. (Possibly because there are often no backup goalie numbers to compare to). So it's likely that Brodeur's playoff GVT underestimates him in this area, along with other strong puckhandlers, including Belfour.
If you look at two-year peak, regular season and playoffs, Bernie Parent's 73-74 and 74-75 are neck-and-neck with Dominik Hasek's 97-98 and 98-99.
Edit: I did a little more digging in the file. It appears that recorded SV% from 1955-67 is used, but not from 1968-82, or pre-1955. This means that GVT may underestimate Bernie Parent's peak! It estimates his SV% was 0.927 in 73-74 and 0.918 in 74-75. In fact, his recorded SV% were 0.933 and 0.924 for those years. Although I'm not sure what GVT estimated for the league averages, it's possible it was low on everyone's SV%.