(An attempt at) The top 40 TPA players from leaderboard scores

[invitro]

Here is another idea I am comtemplating. I might sometime, I don't know when but probably not soon, decide to make high scores on different tables be worth different amounts. Like, a #10 score on a table whose top 100 scores include 40 of the top 100 players, would be worth more than a #10 score on a table with 15 of the top 100 players.

I would do this using the IFPA method of valuing tournaments. (So if doing this is a good idea, then it may add a plus to using IFPA's WPPR ranking points formula.)

This would be a systematic way of doing something I've wanted to do anyway: make TotAN worth less, or maybe nothing at all. And this method might help with the limit-to-23-tables rule, as tables on fewer platforms should have fewer top 100 players. It would also make newer tables worth less, as fewer players have had the time to put a good run on them (should JY be worth the same as AFM? I don't think so).

Any thoughts on this? The first thing to do would be to make a list of the number of top 100 players in the top 100 scores for each table, which I want to do anyway, and will probably do soon.

 

[invitro]

The first thing to do would be to make a list of the number of top 100 players in the top 100 scores for each table, which I want to do anyway, and will probably do soon.

I've done this, and I think the list is very interesting. I expected, maybe, for the best/most-famous tables, TZ / STTNG / AFM, to be on top, and TotAN to be at the bottom, and the least popular tables like Central Park / Victory / Big Shot to be at the bottom. Well, that's not the case... table popularity (or my estimation of it) seems to not be a factor, and recentness of table seems to be a small factor.

Here's the list. The number is the number of top 100 players (calculated by the exp formula; the IFPA formula gives somewhat different results) who have top 100 scores, per table.

43 CV
41 Firep
39 GoinNuts
38 T2
36 Taxi
35 Fl2000
35 Genie
34 MB
34 Gorgar
34 WWind
34 SpSh
33 BRose
33 BigShot
33 NGG
33 Victory
32 TeedOff
32 FT
32 WWater
31 CP
31 TZ
31 BoP
29 HD3E
29 MM
29 ElDor
29 Centaur
29 Pin*Bot
28 SS
28 Co1812
28 RBioN
28 CBW
28 BK2K
27 HH
26 AFM
26 CC
25 FH
24 ToM
23 WD
23 DrDude
23 HS
22 EatPM
21 CentPark
21 STTNG
20 CftBL
20 BH
18 BK
0 TotAN
0 JY

 

[Beatnik-Filmstar]

Thanks for all the hard work crunching numbers that only a handful of nerds will see or care about. It's oddly interesting.

Nice to even crack the list, even if it is in the 80s.

Also, I kinda feel like I should lose points for having my highest placing be on Class of 1812. Rapping and fart sounds are no way to play pinball.

 

[fromduc]

Dont have time now, but if u go to specific problemes on tables, just to let u know that for us PS3 players (and i think but not sure it's the same for PS4 players), our scores on FunHouse and BlackHole are not registered on the leaderboard. At least it was the case until the super patch was applied last week, didnt try after that.

 

[Mark Miwurdz]

Dont have time now, but if u go to specific problemes on tables, just to let u know that for us PS3 players (and i think but not sure it's the same for PS4 players), our scores on FunHouse and BlackHole are not registered on the leaderboard. At least it was the case until the super patch was applied last week, didnt try after that.

My Funhouse top score doesn't appear on the leaderboard when I check using my PS3, but it does when I check using my PS4.

 

[vikingerik]

Also, I kinda feel like I should lose points for having my highest placing be on Class of 1812. Rapping and fart sounds are no way to play pinball.

Protip: turn off the table music in the sound menu and that kills most of the annoying speech as well. You do lose the 1812 theme but I prefer to play without it anyway, it's too hyperactive for a multiball where you need to be deliberately catching and carefully shooting.

 

[vikingerik]

Any thoughts on this? The first thing to do would be to make a list of the number of top 100 players in the top 100 scores for each table, which I want to do anyway, and will probably do soon.

The math sounds intense. Doesn't this end up creating a feedback loop? The point value of each table changes depending on who the top 100 players are... which changes based on the point values of each table. I suppose a closed-form solution is possible but that's talking about solving 100 simultaneous equations which would take a PhD in Wolfram Mathematica or some such.

More seriously, this sounds more trouble than it's worth and makes the chart that much harder to understand. Keep it simple with the tables equal. TOTAN may be worth just singularly ignoring as a special case; there really isn't any other table that has its problems.

Here is how many points this makes several ranks worth (IFPA is the last column):
RANK | linear exp log power desired ifpa |
1 | 100.000 100.000 100.000 100.000 100.000 100.000 |
2 | 99.000 95.456 85.099 95.909 ? 78.162 |
4 | 97.000 86.979 70.198 88.199 ? 57.644 |
8 | 93.000 72.217 55.297 74.513 ? 35.138 |
10 | 91.000 65.803 50.500 68.453 70.000 28.164 |
20 | 81.000 41.333 35.599 44.556 ? 12.192 |
32 | 69.000 23.657 25.495 26.298 ? 9.381 |
33 | 68.000 22.582 24.833 25.152 ? 9.360 |
50 | 51.000 10.244 15.901 11.620 ? 9.020 |
100 | 1.000 1.002 1.000 1.000 1.000 8.020 |
250 | -149.000 0.001 -18.698 0.000 0.100 5.020 |
500 | -399.000 0.000 -33.599 0.000 0.010 0.020 |

So if that's right, the IFPA method slopes downward even more drastically for the top 50 than any of the other methods. Did we want that? It does flatten out nicely below #50 though (and of course it would since it's merely linear from #32 down.)

 

[invitro]

The math sounds intense. Doesn't this end up creating a feedback loop? The point value of each table changes depending on who the top 100 players are... which changes based on the point values of each table.

Yes it does! If it is desired, anyway. I don't know if that Ph.D. is required... there are numerical methods that can solve these things easily and quickly, if the equations have certain properties. The work is in checking that they do. I am not interested in doing that.

More seriously, this sounds more trouble than it's worth and makes the chart that much harder to understand. Keep it simple with the tables equal. TOTAN may be worth just singularly ignoring as a special case; there really isn't any other table that has its problems.

I don't plan to change the table values. Looking at that list was enough.

So if that's right, the IFPA method slopes downward even more drastically for the top 50 than any of the other methods. Did we want that? It does flatten out nicely below #50 though (and of course it would since it's merely linear from #32 down.)

Well, finding out what we want is the difficult part, isn't it? I will keep waiting for feedback, but it is leaning toward just keeping the exp formula with a 23 (half)-table limit per player.

(My main TPA goal has shifted from this to finding a way to get myself re-interested in finishing one of these new Challenges and then finishing all of them.)

 

[JPelter]

I guess the idea was already dumped, but I'd like to pipe up and say that making popular tables worth more points is a terrible idea in my opinion. It'd totally invalidate the efforts of people who like playing EM tables or certain brilliant undervalued gottliebs like tee'd off a lot. Yes I'm looking at my own ass here but I just don't see any real benefit to it at all.

 

[Mark W**a]

I really had to let this sink in before I open my mouth.

The original system, I was happy with.

The system after that, clearly, I was not happy with at all, for valid reasons stated in my old posts.

This system... I feel is the best so far. It truly ranks skill over people who just play a ton of tables.

That said, I want to say I'm cool with either this system or the original.

EXCEPT THAT...I'm actually not cool with that lol. I'm not cool with jumping up 10 ranks, I just feel bad about it.

So, while yes I feel it's the best system of ranking, I think you should just keep it the way it originally was. A vote is a terrible idea, since people are just going to vote what benefits them. Don't over think it, most, if not everyone was fine with how it originally was, right?

 

[Espy]

Just keep it simple. Rank it over all tables. That's a fair enough comparison. After all, we're talking the OVERALL top players.

 

[vikingerik]

The primary objection to using all tables is that not all players have all tables available on their platforms. A secondary one is that some players may be good but uninterested in investing the dozens of hours into every one of the tables. We want to find the best players, not just the ones that have sunk the most time into it.

 

[Kolchak357]

But you also want the best TPA player not just the best AFM player. Not gonna be a good way until all the pins are available on all platforms.

Disclaimer:
I'm just an innocent bystander. I'm probably not in the top 500.

 

[Espy]

Anything other than just plain using the average of every table is going to be an approximation. I don't see it working. It makes the issue far more complex.

If you want to find the best players, run an international tournament or something. There is no way of telling through the scoreboards.

 

[vikingerik]

But you also want the best TPA player not just the best AFM player.

But those are the same thing. (and needs about 10 more smileys)

 

[nuclearkevin]

Gotta give you guys a lot of credit for your math and logic skills at work to figure these methods out! Nice job...

Aside from all the numbers, I would say that if any single player here was the considered to be the best it would most likely be vikingerik. The guy could probably kick everybody's butt on any table he was challenged to.

 

[vikingerik]

Not quite any table. I'm best at tables that have a set strategy to repeatedly grind for points and extra balls. On tables that don't have that (say Central Park or Gorgar), I struggle to make top ten. One could say I'm great at playing Pinball Arcade but merely good at the actual pinball part.

 

[Pinballwiz45b]

Not quite any table. I'm best at tables that have a set strategy to repeatedly grind for points and extra balls. On tables that don't have that (say Central Park or Gorgar), I struggle to make top ten. One could say I'm great at playing Pinball Arcade but merely good at the actual pinball part.

Sorry about what happened on Dr. Dude, by the way

 

[vpalmer]

Heh, Pinkid45b, looks like i kept my promise to beat you in TPA too ~_^

 

[JPelter]

Not quite any table. I'm best at tables that have a set strategy to repeatedly grind for points and extra balls. On tables that don't have that (say Central Park or Gorgar), I struggle to make top ten. One could say I'm great at playing Pinball Arcade but merely good at the actual pinball part.

Don't sell yourself short. You're pretty clearly above everyone else in tenacity, and skill as well. Gorgar does have a clear strategy to grind for points as well, it's just horribly boring.