It’s very hard for me to see how the NYC thing will go away because of the circumstances there where you have a WPPR-loving TD who has a full-time job managing an arcade in a location in one of the largest cities in the world. It’s hard for me to imagine a rule change that can “correctly” nerf any incarnations of the SUPERLeague that don’t hurt 99% of TDs who are doing this in their spare time and putting all the unpaid hours into it.
Regarding using the EFF% for weight in the ranking, I don’t think this is a good idea at all. EFF% is heavily skewed by the number of players in the tournment, due to the dynamic distribution curve. Since SUPERLeague has deliberately tried to get their “regulars” rated so that they have extra padding for the lower spots, you have players (http://www.ifpapinball.com/player.php?p=30136) with very high EFF%, Rating, and Ranking all at the same time because they live in NYC.
Because I’m a masochist . . . I ran the “TrueWPPR” adjustment for everyone in the top 500.
149 of the 500 players participated in more than 1234 points worth of events and get an ‘FLR style’ adjustment for that ‘above and beyond’ play. Francesco actually ends up ranked 108th, and KME moves back to #1 (which is always solid evidence that a change is a good change - funny enough this actually was the test many times in our simulations when Keith was clearly the best player in the land. Any ‘good ideas’ we had were thrown in the garbage when Keith didn’t show up on top of our simulation.
Here’s the top 10 along with a link to the full top 500
1 Keith Elwin 925.39
2 Jorian Engelbrektsson 909.47
3 Zach Sharpe 879.21
4 Daniele Celestino Acciari 824.69
5 Bowen Kerins 755.10
6 Raymond Davidson 718.97
7 Robert Gagno 674.57
8 Jim Belsito 667.69
9 Steven Bowden 663.20
10 Mats Runsten 656.56
For the UK guys:
Martin Ayub - would go from 38th to 34th in the world
Craig Pullen - would go from 72nd to 58th in the world
Greg Mott - would go from 96th to 97th in the world
Peter Blakemore - would go from 99th to 186th in the world
David Mainwaring - would go from 103rd to 253rd in the world
Matt Vince - would go from 123rd to 103rd in the world
Wayne Johns - would go from 331st to 316th in the world
The only reason FLR moves down here is because he does travel to other events. If you look at his SUPERLeauge-only EFF % it’s much higher than his real EFF %. It’s skewed much lower by his PAPA classics results, his Pinholics Anonymous, expo, and anything else he plays in… Implementing this will just get him to stop playing in these other events and make the issue worse.
Beneficiaries of SUPERLeague EFF bump:
95 Lee Hendelman 293.10 20
265 Craig Cash 186.97 14
213 Frederick Asher 204.09 12
23 Sean Grant 508.74 0
252 Matthew Carlson 190.20 14
Clearly the quick proposal I laid out absolutely murders FLR, because he plays in a ton of ‘real events’ (I’m not saying SUPERLeague isn’t real . . . just . . you know . . .)
Hendelman is always the guy I actually look at because there is this overvalue of skill based on our other metrics (or he’s AWESOME) lol. Looks like he would move from 115th to 95th, and I’m not sure what that means because I don’t his ‘skill level’.
But will those beneficiaries be able to actually jump up into the top 10/25/50 based on just their NYC performance (while on purpose not playing in events that will hurt their efficiency percentage).
Lee has only played in 644 points worth of events, so he would have to play twice is much as that at the same ‘rate’ just to maintain his now 93rd place rank. Lee is NOT getting into the top 50 this way ever, because he’s now battling his inflated Eff % just to be able to keep those additional points earned.
Unfortunately FLR can’t ‘unplay’ tournaments he played in
He could stop playing to help his Eff%, but at some point if he gives up playing in the bigger events, he gives up chasing for the top 50 . . . and maybe that’s a good thing??
Good thing being that if someone wants to not play to protect their Eff%, that’s ON THEM. I don’t feel bad at that point de-motivating people to play beyond a normal amount, especially those cherry picking events where they can keep their Eff% high.
I would suggest that the best secondary stat to look into for adjusting a player’s ranking is Glicko/PARS rating, but not as it is currently implemented.
My best suggestion for testing would be to collect true head-to-head and match play results by standardizing the possible finals formats you can play in, force TDs to upload the actual match play or bracket results. Then use that data for adjusting the player’s ranking points.
Currently Lee has 0 true head-to-head games ever played in an IFPA-sanctioned tournament (because of multiball rule) so the confidence his actual skill would be extremely low. In PPL or another League Manager-run league, there is tons of data available. @haugstrup also has tens of thousands of results somewhere.
I can’t see using EFF %, since that is a function of events themselves, becomming anything other than a circular feedback loop. If I think now that EFF % may be used in the future to adjust my rank retroactively, I would want to quit the league where I missed a few nights for being sick and probably will only be able to make 3rd-5th place, even if it might be enough points to get on my card currently because it could take a bigger % of points off my other events on my card.
Read through it a few times to get my head around it and the data looks good in terms of those people I would expect to move up, moving up. And those moving down, moving down.
Interesting data also into average potential points in the top 500 players resume’s. As you have used me as an example a few times here, 985 points worth of action is definitely falling in the, sorry, get out and play ‘to the average’ to be ranked properly category!! I knew I had slowed down in the last year or so, but not to quite that extent! Clearly I need to play more.
Overall though, the initial taste test of the “TrueWPPR” seems pretty good,
I think you underestimate the ability to get all of this data, regardless of the data that Andreas has (with respect to the events held globally).
This also assumes that I only value head-to-head games as determining “who is more skilled” . . . and I DON’T. My biggest problem with PARS was that qualifying data meant NOTHING towards determining the skill level of a player.
For many tournaments only what 10-25% of the player pool is advancing to some sort of ‘direct play’? That means that for most of those players there is NO DATA.
This then throws match play formats into being the only relevant format with respect to accurately rating the skill level of players, and Best Game/PAPA Qualifying formats aren’t worth anything. I think there’s a ton of value in the qualifying process of those tournaments as absolutely showing who is the more skilled/less skilled player.
The majority players moving up are mostly players who are unaffected, and move up to fill the catastrophic fall of the player in this simulation who plays more often.
This may be true, but I do think that the current rating method is a shortcut at best, and appropriately not the sole determination of ranking. Eff % is a fun stat but I think an awful one to try to incorporate into ranking in any way. I see a potential for “cost of living adjustment” but not this one.
Of course you ran this right as Pinvasion Main but not Classics results came in.
It really ends up requiring deeper analysis because I do think looking at Eff% is still a valuable metric in the story of explaining the skill level of a player (it leads to the discussion that they are over/under ranked based on X,Y,Z).
Ultimately something like this might be the best version of a “cost of living adjustment”. Remember that 70% of the players in the top 500 would see NO ADJUSTMENT because they aren’t playing above the average level of events. Looking at that number, 10% more of those players would benefit from their ‘above average’ Eff% against the rest of the top 500 field.
This leaves 20% of the top 500 that play in more than the average number of events, and play them at an efficiency rate that is less than the top 500 average. At some point is it fair to say that those players are over-ranked based on their level of play with respect to their top 500 peers?
If this “cost of living adjustment” isn’t based on your Eff%, it could also be a fixed % . . . any points earned over the average expected level of play for a player will be counted at 50% value. The only people impacted there are the 10 players in the world that actually average more than a 50% WPPR take at the events they go to (amazing).
Right now, as someone who has only played for 3 years, EFF % is the stat that lags behind everything else. This is because 3 years ago I sucked because I had never played pinball before. I don’t like the idea at all that if I win Pinfest I only effectively get 34 WPPRs, but if Trent wins he gets 45 because he was also a top 25 player 3 years ago when I first got ranked at 18,000 and played about at that level for a while.
So, if you were to make this change, and I was above the cut line of available points (currently set at 1234) I would quit Rock Fantasy league, quit Orange County club, and only play Superleague where I now have an EFF % cushion as well. And I sort of have to ride out my 2013 and 2014 garbage.
I do think my results in the past 8 months are pretty good! Before that, not so much…
This would certainly involve a change to the Eff% (which right now COULD actually been on your entire history of data - not just active). Brian thought it was for ‘all time’ but we haven’t looked at the calculation in forever because it’s just a fun stat.
Assuming it is for all of time, then changing it to be ‘active events only’ would be a good change. If it’s ‘active events only’ then a change to ‘last 12 months’ could be a better gauge (knowing that come January 1st of the given year - those additional points earned are based on your Eff% of the previous full calendar year).
Remember there’s another wrinkle in all of this, and that’s AMOUNT OF WORK NEEDED TO MAKE WPPR CHANGES.
Switching to version 5.0 pushed the limits of Shepherd burn out, because that was an insane change from 4.0. This kind of proposal of a “cost of living adjustment” based on “something” (whether it be Rating/Eff%/Fixed value) I believe would be in the doable wheelhouse . . . but even this might be too much of a pain in the ass to implement.
Considering how hard it is to wrangle results out of TDs right now I cannot imagine how it would ever work to get head-to-head results from all tournaments.
I mean, if TDs can’t even be bothered to type of a list of player names in the order of finish and hit a button…
