|
Post by Glenn-Philadelphia on Dec 31, 2022 11:33:24 GMT -5
Very early review of the shot super stats after 15 games.
Here are the on goal pct from shots taking from each zone on the offensive side of the ice:
Z15: 0.527 Z14: 0.622 Z13: 0.541 Z12: 0.427 Z11: 0.450 Z10: 0.429 Z09: 0.333 Z08: 0.398 Z07: 0.330 Z06: 0.260 Z05: 0.241 Z04: 0.265 Z03: 0.341 Z02: 0.174 Z01: 0.236
Here is the save percentage from each zone as well
Z15: 0.861 Z14: 0.893 Z15: 0.921 Z12: 0.915 Z11: 0.825 Z10: 0.887 Z09: 0.948 Z08: 0.925 Z07: 0.918 Z06: 0.930 Z05: 0.929 Z04: 0.910 Z03: 0.933 Z02: 0.875 Z01: 1.000
Note, after reviewing the numbers on the save percentage query, I looked at the data and saw that players from the slot (zone 11) had a higher percent chance to score than from the crease. I adjusted this downward accordingly (those changes will take effect moving forward starting with the games tonight).
I am working through my work/work year end closing today and over the next few days but I plan on updating these stats after a few days and replying in more detail to the questions brought up over the past few days.
The bottom line is we are all looking to make the sim as realistic as possible so as any ideas on how to improve ratings for players are appreciated.
|
|
|
Post by Dane-Hamilton on Jan 2, 2023 12:47:15 GMT -5
hmm, weird pattern was 15/13 and 12/10 should be similar, same zone just opposite side. Looking at it the on goal pct are almost identical but then we get .861/.921 and .915/.887 for sv perc.
Thats a stark difference in sht perc despite being identical opposites. Could just be the low sample #s though.
|
|
|
Post by Chris-Suffolk on Jan 2, 2023 15:49:00 GMT -5
hmm, weird pattern was 15/13 and 12/10 should be similar, same zone just opposite side. Looking at it the on goal pct are almost identical but then we get .861/.921 and .915/.887 for sv perc. Thats a stark difference in sht perc despite being identical opposites. Could just be the low sample #s though. This is an interesting find Dane. Could it be because there are predominantly Left Hand catching goalies and the sim is allowing more blocker side goals than glove side goals? (Based upon the majority of goalies being Left Hand Catch/Right Side Blocker)
|
|
|
Post by Jon-Seattle on Jan 2, 2023 16:27:53 GMT -5
I actually think the difference makes some sense if we are seeing a difference on the NHL level. This would likely be from mirrored positions having more or less shots/goals on one side due to handedness of the shooter and yes glove vs blocker side of the goalie with a significant majority of goalies and skaters being R hand dominant. That said I’m not sure if the save% accurately reflects that without a ton of more information, just that I can see this lining up correctly given the proper circumstances that not even I am comfortable diving into. It was surprising to see though as initially when I was doing my research yesterday I had grouped up quite a few of the positions on the ice initially not thinking there was such a distinct difference particularly between mirrored portions of the ice.
|
|
|
Post by Glenn-Philadelphia on Jan 3, 2023 14:02:01 GMT -5
hmm, weird pattern was 15/13 and 12/10 should be similar, same zone just opposite side. Looking at it the on goal pct are almost identical but then we get .861/.921 and .915/.887 for sv perc. Thats a stark difference in sht perc despite being identical opposites. Could just be the low sample #s though. I am attributing this to a smaller sample size right now. We will see how this looks on future extracts.
|
|
|
Post by Glenn-Philadelphia on Jan 3, 2023 14:04:51 GMT -5
I actually think the difference makes some sense if we are seeing a difference on the NHL level. This would likely be from mirrored positions having more or less shots/goals on one side due to handedness of the shooter and yes glove vs blocker side of the goalie with a significant majority of goalies and skaters being R hand dominant. That said I’m not sure if the save% accurately reflects that without a ton of more information, just that I can see this lining up correctly given the proper circumstances that not even I am comfortable diving into. It was surprising to see though as initially when I was doing my research yesterday I had grouped up quite a few of the positions on the ice initially not thinking there was such a distinct difference particularly between mirrored portions of the ice. As the sample size builds, we will be able to mine this data for each skater/goalie/team too to determine what the numbers are for each of these entities. I might be able to put a page together that would allow someone to dig pretty deeply into these options and not have to rely on me to run a sql query to drop the data onto the forums here. This will become more useful as the database becomes more populated.
|
|
|
Post by Glenn-Philadelphia on Jan 7, 2023 11:59:57 GMT -5
Updated breakdown after 65 total games below. Note, this still contains the elevated values for zone 11. I have also added another column showing total shots on goal from each of the offensive zones.
Now that most of my year end activities are nearly completed, I will be able to dig a little deeper into these as I believe our GPG across the board is still 20-25% higher than it should be.
SOG Pct Z15: 53.40% Z14: 60.30% Z13: 53.80% Z12: 40.50% Z11: 49.70% Z10: 40.50% Z09: 31.80% Z08: 39.10% Z07: 32.10% Z06: 25.50% Z05: 30.80% Z04: 29.40% Z03: 26.70% Z02: 25.70% Z01: 20.30% Goalie Save Pct Z15: 0.929 Z14: 0.823 Z13: 0.931 Z12: 0.89 Z11: 0.883 Z10: 0.911 Z09: 0.93 Z08: 0.927 Z07: 0.928 Z06: 0.938 Z05: 0.922 Z04: 0.937 Z03: 0.981 Z02: 0.982 Z01: 1 Shots on Goal Z15: 622 Z14: 802 Z13: 666 Z12: 382 Z11: 639 Z10: 381 Z09: 228 Z08: 331 Z07: 249 Z06: 225 Z05: 295 Z04: 284 Z03: 52 Z02: 55 Z01: 48
|
|
|
Post by Jon-Seattle on Jan 7, 2023 17:33:54 GMT -5
Good stuff. So is the shots on goal before or after the shot on goal %? I’m pretty sure I know what you were going for on this but I wanted to be sure. So if I’m understanding this correct out of all the shots ~60% actually are on net from zone 14 and that total was 802. The average (or initial/before further adjustment) goalie save % should be 0.823. So with rates evening out for all good/bad Goalie (with the same for the shooter) we should see around 142 goals from zone 14 during that time? I know that’s unlikely to be the exact number but should be somewhere in that range unless significantly more calculations are involved which would obviously give the possibility for wider range or push the number more +/-.
One of the things I’ve noticed is SOG stats are significantly higher in the GHL than NHL. Last year league leader was Austin Matthew’s 345 while halfway through our season he has 285 and on pace for 500+. Now if this is before the calculation of shot actually being “on net” than I think the shot calculation is too low and even more so for guys with lower shot stats as I have guys like Tanner Pearson who had 159 shots over 68 NHL games while in the GHL hes showing 42 shots in 40 games. This is obviously anecdotal and you can all tell me it’s the situation but it’s something I noticed last night across my lineup which makes me believe that shot volume might need adjusting to even them out. Sure merging like spreading out more shot chances while increasing chance of good players scoring by a slightly smaller amount (as our scoring numbers seem slightly high but for the most part accurately represented on the upper end, but drops off considerably faster than it should). Does that make sense or am I way off base here?
Edit: I removed the reference to an earlier thread as the stats were for a separate period and later realized we could ID total goal # and per zone numbers given these stats as long as they are what I think they are.
|
|
|
Post by Jon-Seattle on Jan 8, 2023 0:03:46 GMT -5
So the math once again isn’t adding up. This would suggest that over 65 games there were 498 goals (using SOG * (1 - save%)) per zone. Earlier it was stated that through the first 56 games there were 520 goals with an average of 7+/game after that point. That should put us somewhere around 567 rather than the 498 that these numbers suggest. Even with empty net goals I don’t see the goal differential adding up. I could also be missing other factors? Such as, these numbers suggest we should have seen 81 shots per game.
That said, during this time period Z14 only accounted for 29% of goals. You mentioned changing Z11 to a lower % chance to score but regardless that shouldn’t affect the other areas save % right? That would mean that save %, and overall and on average, should go up from here. This shows average total save % should be .905 (again variation based on goalie/team/etc). That said if these stats are correct then the shot #s and spread should still be accurate and give us a good idea for what the %s should be for where we should be seeing shots from.
Zone shot % Z1 - 1 Z2 - 1 Z3 - 1 Z4 - 5 Z5 - 6 Z6 - 4 Z7 - 5 Z8 - 6 Z9 - 4 Z10 - 7 Z11 - 12 Z12 - 7 Z13 - 13 Z14 - 15 Z15 - 12
And if I say take these numbers over and apply them to my goalie I get a few funky numbers. like a save % of .740 in z14 (with around 138 shots) and an .810 in Z9 but without the actual numbers it’s hard to say how accurate that is. (Especially considering another scenario a few days ago shesty could be seeing as many as 338 shots from zone 14, or is seeing a save % as low as .740, but likely somewhere in between which means he seems to be getting absolutely shafted somehow.)
|
|
|
Post by Glenn-Philadelphia on Jan 8, 2023 10:04:30 GMT -5
I just double checked the figures and everything looks good. The 520 you referenced before was from this post gtgfhl.proboards.com/post/54914/threadwhere I said that there were 520 goals in the first few weeks of play at which point the tables were adjusted to lower the scoring. Those 520 goals are not part of the analysis because detailed shot information wasn't being tracked at the start of the season. All of the data that we are reviewing now is only looking at games that were played since the shot superstats feature was activated and the data was being collected from each game. I just double checked save percentage from each column so there is no need to try to back into this stat as both the shots on goal and the goals are being tracked. The 60% from zone 14 means that of all of the shots attempted from that zone, 60% are on net (not wide or blocked). To determine the percentage of shots from that zone, you would take the actual shots on goal from zone 14 (802 at the last reporting) and divide that by the sum of all shots on goal (5,259) as of last reporting to arrive at a percentage of 15.25%. This means that of all shots on goal for these last 65 games, 15.3% came from zone 14. Another thing I was thinking about as I was verifying the shot stats this AM was that there is no provision to weed out empty net shots in this analysis. Lastly, globally applying these percentages to one particular goalie would not be a beneficial analysis. As I stated earlier, once we have a bigger database to draw from, individual goalies can be queried to see what their actual stats are.
|
|
|
Post by Jon-Seattle on Jan 8, 2023 16:20:39 GMT -5
So in other words: 1. My shot information is correct and all the numbers and calculations I have done are correct given the information you provided. 2. Yesterday - "Updated breakdown after 65 total games below. Note, this still contains the elevated values for zone 11." Previous Thread - "GHL Games prior to 10/19 (before fix 80's table scoring tweak below) GP 56 GS 520 GPG 9.29" Today - "All of the data that we are reviewing now is only looking at games that were played since the shot superstats feature was activated and the data was being collected from each game" (all emphasis mine)So is it including those 56 games or not? Maybe it's just a portion and I understand that. Even if every calculation you have provided to us is from a different set of games (the 56 gp sample size, 65 game sample size, year long, etc.) it doesnt matter. 3. All of the calculations I have thrown out here are correct given the information you have provided to us time and again. It does not explain what I am seeing and have shown for specific goalies. please run the numbers for specific goalies and compare them to what we have shown here. something is off. Please do it before the season is over and we cant go back and fix it, at least from this time on. If its entirely correct then please show me the numbers. Explain to me what I am missing.According to this data:Statistic | League Average | Shesterkin | shots per game | 40 | 33 | Save % | .905 | .907 | GAA | 3.77 | 3.07 | Z14 goal % | 29-34% | 42% | Z14 shot % | 15 | ? | Z14 Save % | .823-.893 | ? |
If I have the best goalie with absolutely no sub par stats, seeing less shots than an average goalie, than why is he giving up a higher percentage of goals from the zone with the most shots league wide. Even if he was seeing a significantly higher number of shots against him, if he were even just league average at stopping shots in Z14 (at the worst possible save % we were given of .823) that would amount to him seeing over 25% of his shots coming from Z14. Given his card he should be leagues better than the average goalie so is it even higher? If it is not than we have an even bigger problem because seeing as how this is the statistically highest place to see a shot from on the ice, the statistical probability that your calculations are being implemented correctly drop to near 0. So either my team is just allowing everyone to move to the crease unabated for shots in which case we can move this discussion from goalies to puck movement tracking or we can identify which calculations are being implemented incorrectly.
For the record my money is on the calculations for Z14 to be implemented backwards where good goalies are being penalized by either a mixup in the table or reference lookup calculation. If it isn't that then there seems to be a pass vs shot calculation issue farther upstream. Another interesting case...Frederik Andersen. Team is doing really well, he has a good not great GAA and Save% despite having second best card. Seems to have a ton of goals from that Z14 area at first glance...
|
|
|
Post by Glenn-Philadelphia on Jan 9, 2023 16:07:27 GMT -5
1) I ran a query to tally all of the goals for the games that we now have super shot stats for and the total of the games played matched the total in the goals in the super shots data.
2) Those 56 games (at the beginning of the season) are not included because I hadn't written the code to capture each and every shot location for every game. I was only capturing the location when a goal was scored or when a penalty was called/missed. Which games certainly do matter for this analysis as there have been 2 adjustments noted to rein in the scoring. The stats provided in the detailed shot breakdown are for the most part (aside from a portion of these from zone 11) a reflection of the current engine. Perhaps this is where the confusion lies.
3) I ran the query for Shesterkin and he has only played 2 games since we started capturing this enhanced data. From that data we can see the following.
Zone and SOG:GA
Z01 - 0:0 Z02 - 0:0 Z03 - 1:0 Z04 - 2:0 Z05 - 4:1 .750 save pct Z06 - 2:0 Z07 - 4:0 Z08 - 3:0 Z09 - 4:0 Z10 - 4:0 Z11 - 12:0 Z12 - 2:0 Z13 - 5:0 Z14 - 7:2 .714 save pct Z15 - 6:0
He also had 1 shot from Zone -1 which he saved.
That being said, there is not a statistically valid way to have any precision in extrapolating Shesterkin's stats from the summary data for all goalies that I have provided in the past (prior to this post)
Any changes are going to impact all goalies as there is only one master table that is referenced to determine what happens when a shot is attempted.
As we gather more data a better analysis can be done to determine the best way to lower goals across all games.
|
|
|
Post by Jon-Seattle on Jan 9, 2023 18:22:56 GMT -5
Ok so we had a small adjustment to the scoring algorithm which was separate from some other adjustment and the addition of the zone shot total grab that you put in place. Understood.
I have 32 games of goal data vs several data pulls of different time periods worth of games that suggest there is a strong trend in what I am suggesting. I understand small sample sizes are going to have higher likelihoods of an outlier and stronger ones at that. But year in and year out our best goaltenders have been average performers. I've finally narrowed down the causes a bit, I've shown you time and again that there seems to be a correlation between goals in the zones being off particularly for goalies with strong cards and everything you have pulled seems to corroborate this. What else do you need to show that there is a calculation error going on somewhere? I'm not asking you to revamp your entire table. I'm asking you to ensure that the calculations being utilized are correct and/or that the table is being referenced correctly, especially when a goalie is supposed to have an advantage over the shooter.
There is nothing above that shows that Shesterkin seems to be seeing an abnormal number of shots from Z14 that would suggest his 42% of goals from Z14 makes sense. You can track it as long as you want. I've already shown you the problem. I've spent two years dealing with this, first with Nedeljkovic last year having the what, third best card or something. And again this year with Shesterkin having the best goalie card. Both ended up being average at best. When was the last time our best goalies were at the top of our Save% and GAA leaderboards?
I know we're not going to go back and rerun the season but it'd be nice if we could have the problem fixed so that my team is actually playing on a level playing field. Right now though it seems like we'll just keep playing with one hand tied behind Shesterkins back and watch my team once again be completed shafted. I'd ask for at least an apology if/when you finally diagnose and fix this since my team has been so heavily affected for two years in a row and an acknowledgement that my teams opportunities at being even halfway decent, that I've spent literal years working to build, are being destroyed by bad code. I mean its not like my team started on even footing to begin with but hey as long as I keep paying my dues who cares right?
|
|
|
Post by Glenn-Philadelphia on Jan 12, 2023 10:26:15 GMT -5
Stats after 99 games (games through 1/11/2023)
SOG Pct (of all shots from zone, how many were on net) Z15: 52.80% Z14: 60.90% Z13: 54.80% Z12: 40.90% Z11: 50.10% Z10: 41.40% Z09: 32.20% Z08: 39.40% Z07: 32.60% Z06: 26.90% Z05: 30.40% Z04: 29.70% Z03: 25.20% Z02: 25.10% Z01: 22.30% Goalie Save Pct (Save pct by zone) Z15: 0.932 Z14: 0.822 Z13: 0.929 Z12: 0.897 Z11: 0.879 Z10: 0.913 Z09: 0.923 Z08: 0.928 Z07: 0.935 Z06: 0.945 Z05: 0.921 Z04: 0.934 Z03: 0.974 Z02: 0.987 Z01: 0.988 Shots on Goal (Total shots on goal by zone) Z15: 937 Z14: 1204 Z13: 1027 Z12: 591 Z11: 994 Z10: 600 Z09: 351 Z08: 501 Z07: 372 Z06: 364 Z05: 430 Z04: 427 Z03: 77 Z02: 79 Z01: 82 TOT: 8036 SOG % by Zone (from table above, % of shots by zone) Z15: 11.7% Z14: 15.0% Z13: 12.8% Z12: 7.4% Z11: 12.4% Z10: 7.5% Z09: 4.4% Z08: 6.2% Z07: 4.6% Z06: 4.5% Z05: 5.4% Z04: 5.3% Z03: 1.0% Z02: 1.0% Z01: 1.0%
|
|
|
Post by Dane-Hamilton on Jan 12, 2023 10:37:18 GMT -5
Looking good, the gap between 15/13 and 12/10 seems to have closed with a larger sample size
|
|