|
|
Post by Oxford United on Apr 19, 2024 21:38:04 GMT -7
Here is the updated list of Positions and Shifts, current to May 1, 2024. There were many changes. POS_and_SHIFTS.xls (83 KB) You will notice that only CHL, Rookie and Free Agent players populate the list. That is because I was forced to do a manual correction for each player and didn't want to do that for all of the players not in the CHL. Ach, mein carpal tunnel! - Steve
Attached is an XLS file with *DRAFT* player name, final team, games played, primary position, position rating and shifts rating for every player who played 10 or more games. If you spot any errors, let me know.
(Updated April 29)
Offical Changes since first *DRAFT* version:
Method for calculating shifts changed from MROUND to CEILING. (April 28)Tavares updated from LW to C/LW (April 25)Wahlstrom traded back from EDM to NYI. (April 25)Several defencemen updated from LD to D: Rasmus Dahlin, Cam Fowler, Miro Heiskanen, Travis Sanheim, Vince Dunn, Sean Durzi, Owen Power, MacKenzie Weegar, Jakob Chychrun, Cam York, Mikhail Sergachev, Shea Theodore (April 29)
|
|
|
|
Post by nash on Apr 19, 2024 21:52:28 GMT -7
Wahlstrom classified as being on EDM. (FIXED)
|
|
|
|
Post by Shamrockville on Apr 19, 2024 22:25:59 GMT -7
Just a quick question on one of my guys, Nichushkin played heavy minutes on LW at the start of the season, did he miss W rating due to his minutes played?
And no C rating for Tavares? (FIXED. Correct Rating is C/LW)
|
|
|
|
Post by QC Mike on Apr 21, 2024 11:33:09 GMT -7
What kind of time should we be thinking of for alternate ratings? Was RNH's centering the third line for a dozen+ games to start the season not enough?
|
|
|
|
Post by Shamrockville on Apr 21, 2024 11:45:07 GMT -7
On this line of statistics....I tried to find any source that shows a players time spent playing at any given position. I couldn't find anything anywhere that broke down ice time to that degree. Does anyone have a site or link that gives any of that kind of breakdown? It'd be nice to be able to at least see what kind of ice time deployment a guy has during the season to give coaches even a basic idea of that kind of info.
|
|
|
|
Post by QC Mike on Apr 21, 2024 13:54:52 GMT -7
On this line of statistics....I tried to find any source that shows a players time spent playing at any given position. I couldn't find anything anywhere that broke down ice time to that degree. Does anyone have a site or link that gives any of that kind of breakdown? It'd be nice to be able to at least see what kind of ice time deployment a guy has during the season to give coaches even a basic idea of that kind of info. I had found one paper that broke this down for the Rangers or the Devils for one season (2016 I think) so not much use.
Beyond that Hockey Reference will let you look at an individual's game stats-line for the season but no positional info. Might be able to infer some positional info from it (multiple face offs, probably playing center but for how much of the game, ?).
I think Cap Friendly has their positioning based on the player appearances on depth charts but don't know how many games it takes to earn multiple positions (is a single appearance as a C enough to get that or is it some ratio?).
NHL.com also gets you access to each games individual stats and the lineup (which doesn't mean much as Edmonton has RNH as a C but for the game I looked at he was starting on McDavid's wing).
Dobber gives common line combos by quarter with no ice time breakdown (according to Dobber Kane was centering that third line with RNH taking the faceoffs based on the numbers).
But as for something that says Joe Blow played X amount of time as a RW in the game and Y amount on the left I haven't found anything yet.
|
|
|
|
Post by Oxford United on Apr 21, 2024 16:44:54 GMT -7
The technical insight: players more than one standard deviation over average ice time have one position removed from their ratings.
The non-technical insight: guys who play more play fewer postions.
|
|
|
|
Post by bonavista on Apr 24, 2024 7:48:48 GMT -7
I have a question about shifts. in looking at this and comparing the shifts to what was pulled in using AutoStatz i have some players (13) who lost half a shift. is this intended or is their a rounding error somewhere (on either spreadsheet)? (FIXED. Shifts now rounded up to next 0.5.)
|
|
|
|
Post by bonavista on Apr 27, 2024 22:10:50 GMT -7
Is there a problem with D positioning on Vegas? after watching tonights game and seeing Theodore play the right side i was wondering about his LD rating this coming year.
looking at the ratings there are only 3 players rated RD
Pietrangelo 64 games
Whitecloud 61 games
Korczak 26 games
that is missing 95 games worth of RD. i can see some be rounding errors removing some but not that many.
i think Theodore should probably by RD not LD [FIXED. Theodore is now D]
and then probably Hutton maybe being D (not sure on that but he seems the most likely case to have swapped)
|
|
|
|
Post by Oxford United on Apr 28, 2024 10:10:18 GMT -7
Vegas had 5 skaters rated RD as primary position: Whitecloud, Korszak, Pietrangelo, Miromanov (Traded), Pachal (Traded). - Combined they equal 2.10 RD per Game.
Vegas had 6 skaters rated LD as primary position: Hanifin, Hague, Hutton, Martinez, McNabb, Theodore. - Combined ehy equal 3.87 LD per Game.
Total number of defencemen per game? 5.96. A sliver off 6 defencemen per game.
So there are enough games. Theodore does play LD and RD, though whether he plays one or the other better, I don't know. I do know that LD is is primary position. Theodore likely did/does fill in on RD. Others may have as well.
|
|
|
|
Post by Shamrockville on Apr 28, 2024 19:07:12 GMT -7
Steve, can you explain a bit on this? "Theodore at 11.5 shifts is at or over the "position limit" line, and so had one position stripped from him."
I'm trying to understand if a guy who plays 22 minutes a game, but spends say 12 of those at LD and 10 at RD, would NOT get a D rating? And what is the causation of that if so?
We went through a pretty big change when we brought back in the LD and RD fixed positions so if guys are now, not getting these multi ratings for High shifting players, would that be detrimental to the previous change?
|
|
|
|
Post by nash on Apr 28, 2024 19:29:03 GMT -7
In the CHL we had LD/RD fixed positions for much longer than we had no positions on D.
|
|
|
|
Post by bonavista on Apr 28, 2024 21:16:23 GMT -7
|
|
|
|
Post by Oxford United on Apr 29, 2024 0:48:56 GMT -7
OK, Theodore is a D.
I’ve removed the cap on all defensemen. LD/RD of any number of shifts are now rated D.
|
|
|
|
Post by Shamrockville on Apr 29, 2024 12:51:00 GMT -7
So, I understand that we as coaches have a vested interest in our players. And also that we as hockey fans have certain expectations when a particular players role comes up.
However, we have all agreed that Steve's factual, statistical model has been the best thing we've seen for the game in, probably decades. We no longer have to guess at the ratings, we no longer have to put up with unfair biases. We get actual real results that ensure we are all playing with players on the same playing field and ratings. The stats and formula he uses have proven to be correct for 99% of the players, barring an unforeseen oversight.
I get that our "thinking" of where a guy plays will probably be different, but he's come up with a formula that works. There HAS to be a cutoff in the formulas and sometimes our guys will fall on a side of that cutoff that we personally aren't fond of, but that shouldn't indicate the formula is incorrect.
We all knew that the positions for higher shifting players was going to be a bit more condensed, as we tried to get players more solidified in their roles and we also wanted to spur some more trading for other than the 2 or 3 GM's who make 80% of the trades.
I'm guilty as well of pining for better results for my players, but if Steve says to me "Hey, I checked Val Nichushkin and he's just above the cutoff line and is rated only RW", I just have to accept that. Or, "Hey I checked John Tavares and there was a small slip there", in that particular instance, he acknowledged that there was a small error in that players instance and it was rectified, but he laid out the justification and the players for the positions on VGK that showed the positional rating was accurate based on the formula that we all are using.
My personal thought is that if we have a large issue, that we should probably speak to him directly, and even if we don't quite agree or like the cutoff, that we accept it and take into consideration the enormous effort he put in to come up with this viable effort to control and replicate statistical accuracy rather than the eyeballing and, quite often, nonsensical ratings that we had with HB.
If we feel that the formula should be tweaked somehow, then maybe work with him directly, see how far off a player was from our thought that the rating should be and what it came out with. He's never been shy about tweaking a formula, nor has he ever been averse to looking at wholesale changes, but we have to approach this from a logical and statistical basis rather than a "my guy should be..." scenario. Not to say that Ryan didn't put forth some information to show where his claim was based, but we can't individually argue about ratings, to the point that a change is made just for the sake of that change.
Think of it like this, if you had a project of this scope that had thousands of hours put in, just to make your friends happy, and then we all complain and argue when we see certain results....how would you feel? Frustrated would be an understatement I think.
Just my 0.5 cents worth.
|
|
