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

[invitro]

Refreshing the leaderboards today?

Yep!

 

[invitro]

Rankings for 2014-05-16

Hi, here are the rankings for 2014-05-16. The last ranking was 2014-04-18. I've settled on a schedule of doing this every four weeks, so the next will be on June 13.

There are several big increases and debuts this month. Which is great! There also appear to be several cases of hacked scores (or glitched scores, how to tell the difference?). I have removed six usernames, one of which I was pointed to by this forum.

I still plan to alter the scoring formula from the exponential decline, which I think is too severe, to something "midway between" that and a linear decline. If anyone wants to give me an explicit formula that they think will be good, please do. Here is the current exp formula:
POINTS = 100 * exp(-0.0465 * (RANK - 1))
and here is the linear formula:
POINTS = 101 - RANK

The requirements are:
- RANK = 1 must give POINTS = 100,
- RANK = 100 must give POINTS = 1.

Happy TPAing.

Rank (previous) Change	Username	Points	# of Top 10 Scores (previous)	# of Top 100 Scores (previous)	# of Top 500 Scores	Best Scores
1 (2) +1	AntonR	2984	28 (17)	45 (44)	45	#1 on NGG, #1 on SpSh, #1 on Centaur, #1 on ElDor
2 (1) -1	vikingerik	2928	27 (29)	39 (38)	40	#1 on MB, #1 on Taxi, #1 on AFM, #1 on T2
3 (3) 0	Tarek Oberdieck	1826	19 (18)	20 (19)	21	#1 on BH, #1 on FH, #1 on Genie, #1 on DrDude
4 (4) 0	Sir_Moovenstein	1496	13 (13)	28 (25)	35	#1 on CV, #2 on AFM, #3 on ElDor, #4 on Genie
5 (13) +8	Toolinit	1420	8 (5)	40 (29)	44	#1 on Fl2000, #2 on HD3E, #3 on HighSpeed, #6 on DrDude
6 (5) -1	FALCO12-QUEBEC	1168	9 (11)	26 (26)	30	#1 on EatPM, #3 on RBioN, #4 on Centaur, #6 on CV
7 (7) 0	Space Tractor	1156	6 (7)	27 (29)	41	#3 on Pin*Bot, #5 on SpSh, #5 on FT, #5 on BRose
8 (6) -2	TOUGE_KILLER	1144	9 (9)	29 (30)	38	#2 on DrDude, #3 on Firep, #3 on WWind, #4 on CBW
9 (8) -1	www.vpforums.org	1029	8 (10)	16 (16)	16	#1 on STTNG, #4 on AFM, #5 on BH, #5 on Genie
10 (9) -1	atoppachar	1026	8 (8)	23 (23)	38	#2 on FT, #4 on Taxi, #4 on GoinNuts, #4 on BRose
Rank (previous) Change	Username	Points	# of Top 10 Scores (previous)	# of Top 100 Scores (previous)	# of Top 500 Scores	Best Scores
11 (11) 0	maestroreese	991	4 (3)	34 (32)	43	#3 on DrDude, #9 on Genie, #9 on SpSh, #10 on CP
12 (10) -2	invitro	924	3 (3)	31 (32)	43	#4 on FH, #9 on AFM, #10 on MM, #11 on BH
13 (12) -1	HUFMUF	900	7 (7)	17 (17)	26	#1 on HD3E, #1 on TeedOff, #1 on Co1812, #2 on TZ
14 (17) +3	dylan_h	842	5 (4)	27 (27)	43	#3 on Co1812, #4 on DrDude, #7 on TeedOff, #7 on FT
15 (15) 0	Howling_Goro	776	8 (8)	11 (11)	20	#2 on STTNG, #2 on Centaur, #2 on T2, #5 on BoP
16 (20) +4	Just this guy	751	5 (4)	20 (17)	24	#3 on WD, #4 on HighSpeed, #6 on BRose, #9 on HD3E
17 (14) -3	pinballchris	746	3 (3)	28 (28)	42	#6 on ElDor, #7 on Co1812, #7 on CBW, #11 on NGG
18 (16) -2	Crusty Booger	730	3 (4)	28 (28)	44	#5 on TeedOff, #8 on SpSh, #8 on Co1812, #11 on Victory
19 (22) +3	phoerber	688	7 (6)	10 (8)	10	#2 on BK2K, #2 on HighSpeed, #3 on EatPM, #5 on CftBL
20 (18) -2	geohoo	680	2 (4)	24 (23)	38	#6 on WWind, #9 on CC, #11 on FH, #12 on BK2K
Rank (previous) Change	Username	Points	# of Top 10 Scores (previous)	# of Top 100 Scores (previous)	# of Top 500 Scores	Best Scores
21 (19) -2	TRLBrony	662	3 (2)	25 (25)	37	#6 on Co1812, #10 on Fl2000, #10 on HighSpeed, #15 on CBW
22 (38) +16	ChristofferB	632	4 (2)	23 (19)	38	#9 on GoinNuts, #9 on WD, #9 on HighSpeed, #10 on Victory
23 (21) -2	pinballwiz45b	611	1 (4)	20 (17)	36	#9 on Victory, #11 on Co1812, #11 on FT, #13 on Pin*Bot
24 (24) 0	djchrille	523	3 (3)	13 (13)	23	#5 on T2, #7 on Victory, #10 on HD3E, #11 on BigShot
25 (23) -2	benika7	518	4 (4)	11 (12)	12	#3 on CV, #3 on MB, #7 on CC, #9 on TZ
26 (26) 0	Mark_Miwurdz	513	5 (5)	9 (9)	13	#3 on BoP, #5 on RBioN, #7 on T2, #8 on FH
27 (25) -2	vpalmer	496	2 (4)	16 (16)	26	#7 on SpSh, #7 on Centaur, #11 on Genie, #12 on FT
28 (100) +72	Fromduc	471	0 (0)	20 (7)	31	#14 on SpSh, #14 on CBW, #17 on WWater, #17 on Pin*Bot
29 (27) -2	Phreaker47	456	5 (5)	6 (7)	15	#3 on Fl2000, #3 on HH, #5 on Taxi, #6 on BoP
30 (168) +138	NeonZwei	437	4 (1)	7 (4)	9	#1 on BRose, #3 on AFM, #3 on FT, #6 on ToM
Rank (previous) Change	Username	Points	# of Top 10 Scores (previous)	# of Top 100 Scores (previous)	# of Top 500 Scores	Best Scores
31 (28) -3	PinHead970	435	4 (4)	8 (8)	10	#1 on WWind, #5 on STTNG, #5 on Co1812, #7 on TZ
32 (29) -3	ostrich2	410	4 (4)	13 (13)	17	#2 on CC, #8 on MB, #8 on HD3E, #8 on NGG
33 (5087) +5054	Captain_Cockup	404	2 (0)	12 (0)	16	#6 on FH, #7 on RBioN, #11 on STTNG, #18 on BH
34 (33) -1	efbelo	397	1 (2)	10 (10)	30	#10 on CBW, #11 on MB, #11 on T2, #15 on CentPark
35 (31) -4	moldygoat	372	1 (1)	10 (12)	26	#6 on CP, #12 on WWater, #14 on BoP, #14 on WWind
36 (36) 0	Lupo23	364	3 (3)	5 (5)	6	#1 on CentPark, #2 on BK, #5 on Gorgar, #12 on GoinNuts
37 (40) +3	Tripple I	363	0 (0)	18 (16)	37	#16 on WWind, #17 on SpSh, #21 on CC, #21 on BRose
38 (30) -8	Jake Alt	363	3 (3)	10 (10)	30	#3 on BRose, #4 on WD, #7 on BK2K, #22 on CV
39 (32) -7	chk	360	1 (1)	12 (14)	32	#7 on BRose, #12 on Fl2000, #15 on FH, #15 on BK2K
40 (43) +3	Nemesis041	358	1 (1)	16 (15)	25	#8 on BK2K, #11 on WWater, #14 on HighSpeed, #26 on SS
Rank (previous) Change	Username	Points	# of Top 10 Scores (previous)	# of Top 100 Scores (previous)	# of Top 500 Scores	Best Scores
41 (87) +46	maezun	351	1 (2)	12 (8)	26	#8 on Pin*Bot, #11 on RBioN, #11 on ElDor, #17 on BRose
42 (35) -7	DeeEff	343	0 (0)	14 (14)	39	#15 on WWind, #17 on Centaur, #21 on TeedOff, #22 on FH
43 (37) -6	CheesyBeefy88	335	1 (1)	18 (18)	34	#8 on BigShot, #15 on SpSh, #18 on CP, #18 on GoinNuts
44 (44) 0	banane32	335	4 (4)	4 (4)	4	#2 on NGG, #3 on MM, #5 on MB, #10 on SS
45 (34) -11	Nightmac21	330	0 (0)	18 (19)	36	#12 on TeedOff, #20 on CV, #24 on HD3E, #26 on CBW
46 (39) -7	PieCES	322	1 (1)	17 (18)	31	#10 on Firep, #12 on BH, #12 on DrDude, #32 on CV
47 (42) -5	rockman_x_2002	321	2 (2)	12 (12)	31	#6 on GoinNuts, #6 on HH, #17 on Fl2000, #28 on Firep
48 (45) -3	Chopin6868	313	2 (2)	12 (13)	23	#2 on WWind, #10 on Pin*Bot, #15 on HH, #19 on BoP
49 (41) -8	ParallaxScroll	312	0 (0)	12 (12)	20	#17 on MM, #18 on CV, #19 on Gorgar, #24 on BigShot
50 (61) +11	Jazza	305	1 (1)	19 (13)	40	#9 on Firep, #20 on CentPark, #22 on CBW, #28 on TeedOff
Rank (previous) Change	Username	Points	# of Top 10 Scores (previous)	# of Top 100 Scores (previous)	# of Top 500 Scores	Best Scores
51 (46) -5	MT	298	0 (0)	18 (18)	40	#14 on TeedOff, #20 on Firep, #23 on WWater, #26 on RBioN
52 (52) 0	II-Squid-II	294	1 (1)	8 (8)	13	#1 on RBioN, #11 on BoP, #15 on CV, #19 on MB
53 (50) -3	kovalev1985	294	2 (2)	8 (9)	18	#2 on CV, #8 on Taxi, #19 on RBioN, #26 on Centaur
54 (54) 0	ER777	294	1 (1)	15 (11)	39	#5 on ElDor, #16 on CP, #19 on Genie, #25 on HH
55 (47) -8	tuomas	290	1 (1)	10 (11)	17	#6 on AFM, #22 on SpSh, #24 on BH, #25 on Firep
56 (51) -5	RIZ420_420_1	286	1 (2)	11 (11)	16	#5 on WWater, #11 on Pin*Bot, #21 on CP, #26 on CC
57 (49) -8	Roy Wils	284	0 (0)	11 (11)	16	#13 on CV, #16 on FH, #19 on HD3E, #27 on AFM
58 (53) -5	LordEngi	281	1 (1)	12 (12)	27	#10 on HH, #13 on CBW, #19 on GoinNuts, #26 on Firep
59 (60) +1	OmegaDef	275	2 (1)	8 (7)	16	#1 on CftBL, #9 on CP, #20 on Genie, #22 on RBioN
60 (48) -12	gooche77	274	1 (1)	14 (13)	29	#7 on Genie, #24 on Firep, #29 on ElDor, #29 on FT
Rank (previous) Change	Username	Points	# of Top 10 Scores (previous)	# of Top 100 Scores (previous)	# of Top 500 Scores	Best Scores
61 (185) +124	Benj341276	265	2 (1)	10 (5)	29	#5 on CC, #6 on MB, #28 on WWater, #32 on Fl2000
62 (57) -5	siemas12345	259	2 (2)	6 (7)	19	#2 on SS, #9 on T2, #14 on HH, #42 on CC
63 (55) -8	Caps_Lock420	254	1 (1)	13 (13)	26	#4 on ElDor, #20 on WWind, #28 on BigShot, #33 on Centaur
64 (62) -2	smbaydp	246	2 (2)	12 (9)	40	#9 on Fl2000, #10 on ElDor, #20 on BigShot, #43 on Taxi
65 (58) -7	sneeking	242	2 (2)	9 (10)	17	#9 on Centaur, #10 on SpSh, #18 on Pin*Bot, #28 on Genie
66 (56) -10	Lordbobby	240	0 (0)	9 (10)	21	#12 on Co1812, #20 on SpSh, #25 on GoinNuts, #26 on Genie
67 (59) -8	tiotbellot	240	0 (0)	8 (8)	19	#11 on TeedOff, #18 on HH, #22 on T2, #23 on WWind
68 (111) +43	ppeltola768	232	1 (1)	13 (10)	32	#9 on NGG, #14 on Victory, #28 on BK2K, #34 on HighSpeed
69 (65) -4	wfusdfcfdedeji83	226	0 (1)	4 (5)	9	#11 on AFM, #12 on RBioN, #13 on MB, #18 on HD3E
70 (63) -7	kefran22	223	0 (0)	8 (8)	15	#12 on Firep, #14 on Genie, #24 on BoP, #24 on CftBL
Rank (previous) Change	Username	Points	# of Top 10 Scores (previous)	# of Top 100 Scores (previous)	# of Top 500 Scores	Best Scores
71 (64) -7	Daniel Rone	221	1 (1)	7 (7)	10	#5 on CV, #14 on BigShot, #30 on ToM, #30 on Firep
72 (70) -2	xNiCeGuYx	218	0 (0)	13 (12)	31	#15 on WD, #19 on Centaur, #20 on BK2K, #24 on GoinNuts
73 (66) -7	ALOL997	217	1 (1)	7 (7)	17	#9 on CV, #18 on RBioN, #23 on EatPM, #28 on Gorgar
74 (69) -5	Mark_Miwurdz_US	216	0 (0)	6 (6)	7	#12 on STTNG, #16 on TeedOff, #17 on T2, #20 on FH
75 (68) -7	Actionball	215	1 (1)	5 (6)	8	#10 on CftBL, #12 on Gorgar, #16 on Taxi, #25 on ToM
76 (71) -5	ThePinbugArcade	212	2 (2)	3 (3)	4	#4 on Firep, #7 on Gorgar, #16 on CentPark, #118 on RBioN
77 (76) -1	Nelis Bakker	208	1 (1)	4 (4)	5	#3 on WWater, #12 on TZ, #20 on BoP, #40 on SS
78 (77) -1	pootV3	208	2 (2)	3 (3)	9	#3 on HD3E, #7 on STTNG, #20 on MB, #144 on TZ
79 (74) -5	krazysteve1958	206	2 (2)	6 (6)	22	#3 on FH, #5 on CBW, #52 on ElDor, #53 on HighSpeed
80 (67) -13	olivierarmy	200	0 (0)	12 (12)	14	#17 on MB, #18 on CC, #20 on CP, #34 on BRose
Rank (previous) Change	Username	Points	# of Top 10 Scores (previous)	# of Top 100 Scores (previous)	# of Top 500 Scores	Best Scores
81 (--) *	West_Tiger	191	2 (--)	5 (--)	12	#2 on Firep, #6 on CentPark, #52 on HighSpeed, #63 on Gorgar
82 (72) -10	yoshida758do	190	0 (0)	8 (8)	13	#20 on WWater, #21 on SS, #22 on CC, #27 on FT
83 (82) -1	roma	189	2 (2)	3 (3)	3	#3 on BH, #3 on Genie, #57 on EatPM
84 (73) -11	nikostos44	189	1 (1)	13 (13)	25	#3 on GoinNuts, #21 on Victory, #39 on SpSh, #47 on BoP
85 (81) -4	switch3flip	187	1 (1)	8 (8)	21	#6 on Victory, #21 on FT, #31 on WWind, #37 on CP
86 (80) -6	redflame2013	187	2 (2)	3 (3)	5	#4 on BH, #6 on Taxi, #35 on BigShot, #107 on BK
87 (84) -3	bobo	186	2 (2)	2 (2)	3	#2 on BH, #3 on Gorgar, #400 on Firep
88 (1126) +1038	Eoner321	185	0 (0)	7 (2)	20	#16 on Firep, #17 on HH, #20 on HighSpeed, #23 on Pin*Bot
89 (79) -10	WIP	183	1 (1)	8 (8)	11	#10 on MB, #15 on AFM, #24 on CP, #39 on FH
90 (85) -5	nuclearkevin	183	2 (2)	3 (3)	4	#4 on TZ, #9 on STTNG, #29 on T2, #176 on BoP
Rank (previous) Change	Username	Points	# of Top 10 Scores (previous)	# of Top 100 Scores (previous)	# of Top 500 Scores	Best Scores
91 (169) +78	FalconPain	179	0 (0)	5 (3)	19	#12 on NGG, #14 on CP, #21 on HD3E, #36 on AFM
92 (643) +551	Dizzer2012	179	0 (0)	7 (2)	16	#12 on CC, #26 on CentPark, #26 on Fl2000, #31 on Pin*Bot
93 (83) -10	evilryu007	178	1 (1)	4 (4)	15	#8 on T2, #16 on CV, #19 on WWater, #48 on SS
94 (241) +147	baelgor	175	0 (0)	5 (2)	9	#14 on MM, #18 on MB, #21 on AFM, #24 on CV
95 (75) -20	slam23	175	0 (0)	15 (14)	35	#19 on ElDor, #33 on FT, #34 on Victory, #37 on MM
96 (78) -18	CentralScrutnizr	173	2 (2)	4 (4)	16	#8 on FT, #10 on BRose, #24 on Pin*Bot, #100 on T2
97 (90) -7	bach211i	172	2 (2)	2 (2)	3	#1 on GoinNuts, #8 on Fl2000, #114 on BH
98 (93) -5	autoteo78	169	0 (0)	4 (4)	8	#11 on HD3E, #16 on TZ, #26 on MB, #31 on SS
99 (97) -2	Zoop25	169	0 (0)	9 (8)	31	#14 on Co1812, #28 on CP, #28 on GoinNuts, #39 on Genie
100 (86) -14	victorstulemeije	168	0 (0)	15 (17)	34	#12 on HD3E, #25 on TeedOff, #36 on Victory, #48 on CV

 

[invitro]

I don't think it's fair to give someone, well, anything for a score that's outside the top 100. Because basically what that does is reward people who just have more tables, vs highly skilled players.

I started out giving 0 pts for scores under #100. There seemed to be a lot of demand to give ranks #101-500 something, so I did... although it is almost zero with the exp formula. I think I completely agree that <#100 should be worth nothing, but a smooth scoring curve seems better. Or maybe I'm just being indecisive.

I am always listening to ideas on formulas. vikingerik had the key idea to make a #1 rank worth lots more than two #50 (or #250 ranks). I think I am still considering making the number of scored tables drop off... like have the best 23 (or half) tables count fully, then count the others less, to compensate some for platforms that don't have the full array of tables.

 

[vikingerik]

Wow, neck and neck on me and AntonR. But the timing just missed some games I just played tonight, took #1 on Harley and #3 on Whirlwind.

As for the dropoff formula, here's some brainstorming. Make it so that every doubling in rank is equal. So the difference between places 1-2 is equal to the difference between places 2-4, 4-8, 8-16, and so on. In other words, every time you improve yourself past half the players above you, that counts equally. Going from #100 to #50 is the same as going from #80 to #40, or #30 to #15, or #4 to #2, or any other such combination. This feels nicely elegant mathematically.

This also has the advantage of scaling out to however far down the leaderboard you want to go while always working the same way. (It would work best going to a power of 2 like 128 or 512 rather than 100 or 500.)

So suppose you want place #1 to have 100 points and to rank down to place #512. There are 9 doublings in that range, since 2^9 = 512. The closed-form expression is (1-log2(RANK)/9) * 100.

Edit: actually, a power of 2 isn't necessary. You can rank down to any place N with this expression: (1-log(RANK)/log(N)) * 100.

 

[Mark Miwurdz]

Are you taking glitched/hacked scores into account? For instance, my true rank on Ripleys would be 5th and 8th on Funhouse are just 2 examples.

 

[OmegaDef]

Thank you so much for taking the time to make this list. Congratulations to everyone who made it!

 

[invitro]

Are you taking glitched/hacked scores into account? For instance, my true rank on Ripleys would be 5th and 8th on Funhouse are just 2 examples.

I am indeed ignoring usernames with a hacked score. I was not doing anything about glitches because I didn't know of them... except for the RBioN one, which I knew about, but hadn't implemented, oops! I am now skipping RBioN scores that match 10,0xx,xxx,xxx or 20,0xx,xxx,xxx. This gives you your #5 rank. The FH issue is a 999,999,999 score that I hadn't skipped; it is skipped now. I've edited the top 100 table (just the May 16 one, not previous ones). This bumps you up to #26 from #27.

Thanks a bunch for the feedback, this is very helpful.

 

[Jazza]

Thanks for your work on this invitro. It adds an element of fun to an already great platform!

 

[invitro]

As for the dropoff formula, here's some brainstorming. Make it so that every doubling in rank is equal.

Thanks for the feedback. This formula would make the difference between #1 and #10 equal to the difference between #10 and #100, though. I require #100 to be worth one point, so #10 would be worth about 50 points, and that's even lower than what happens in the exp formula, where #10 is worth 66 points. (To fix f(1) = 100 and f(100) = 1, I get a formula of f(x) = 100 * (1 - log(x)/log(104.76)), which gives f(10) = 50.5.)

I am unsure that I did that right.

I suppose I should try to be more specific about the values of the desired function f(x). I require f(1) = 100 and f(100) = 1. I think we want f(50) to be worth more than the current exp formula gives, which is 10.2, so maybe f(50) = 20 (I'm guessing). And I want ranks #101-500 to be worth more also, so maybe f(250) = 0.1, and f(500) = 0.01.

Here is a table of various functions. They are:
linear: PTS = 101 - RANK
exp: PTS = 100 * exp(-0.0465 * (RANK - 1))
log: PTS = 100 * (1 - log(RANK)/log(104.76))
power: PTS = 100 * ( (1 - (RANK-1)/499) ^ 20.825)

RANK | linear exp log power desired |
1 | 100.000 100.000 100.000 100.000 100.000 |
2 | 99.000 95.456 85.099 95.909 ? |
4 | 97.000 86.979 70.198 88.199 ? |
8 | 93.000 72.217 55.297 74.513 ? |
10 | 91.000 65.803 50.500 68.453 70.000?? |
20 | 81.000 41.333 35.599 44.556 ? |
50 | 51.000 10.244 15.901 11.620 ? |
100 | 1.000 1.002 1.000 1.000 1.000 |
250 | -149.000 0.001 -18.698 0.000 0.100 |
500 | -399.000 0.000 -33.599 0.000 0.010 |

(I am messing around with the power function.)

Well, I am not sure what to think. Except that it seems that it's required to decide how many points #10 and #50 should be worth.

 

[vikingerik]

I require #100 to be worth one point

Why do you require this? Is there a mathematical basis, or just doing it that way because you have been? All sorts of functions are easier to work out with a baseline of 0 rather than 1. The difference between 0 or 1 point at the bottom of the board won't mean anything for the total rankings and makes it mathematically cleaner.

But anyway, that doesn't really matter. Yeah, what shape to make the curve is a bigger question. I think the log method is naturally emergent from mathematics, while the other non-linear methods are artificially constructed to fit what you/we already decided we wanted the curve to look like. The log method needs no magic-constant besides how deep you want to rank.

Thinking about other methods in the real-world for comparison... NASCAR ranks linearly, one point per race position. Golf ranks on an arbitrary schedule: "major winners get 100 ranking points. The second place golfer gets 60% of this amount, 40% for 3rd, 30% for 4th, 24% for 5th, down to 14% for 10th, 7% for 20th, 3.5% for 40th to 1.5% for 60th." Tennis does about the same as golf with fixed numbers per event rather than percentages.

Hey, how about we use the PAPA ranking method? http://www.ifpapinball.com/ranking-info We can treat each table as a tournament of equal value, skipping that whole complicated section about Tournament Value Adjustment. Then the ranking uses the sum of a linear method through the entire population plus a power method for the top 32 players. Interested in that at all?

 

[ER777]

I think you guys are overcomplicating this. Just give out some bonus points for the top few places, maybe up to top 10, then the liner progression for the rest of the top 100, and half a point for whatever honorable mention slots you want to include up to 200 or whatever.

The number one spot should be worth significant bonus points though, then dropping off for second and third and disappearing by either fifth or tenth. Like 200, 150, 125 for the top three spots respectively would make sense. A number one spot should be worth a lot more than something from the teens or worse, that's the key point here.

 

[invitro]

Here is the top 100 using a log function for scoring ranks. To be specific, it uses

POINTS = 100 * (1 - log(RANK)/log(500))

I have not looked at the differences between these rankings and the ones for the exp function, except for my own, and I think that #7 for me is just too high, and the previous month's #4 is really way too high (the #12 from the exp function seems right).

I'm posting this for comparison; I hope to decide on a function soon.

Rank (previous) Change	Username	Points	# of Top 10 Scores (previous)	# of Top 100 Scores (previous)	# of Top 500 Scores	Best Scores
1 (2) +1	AntonR	3072	28 (17)	45 (44)	45	#1 on NGG, #1 on SpSh, #1 on Centaur, #1 on ElDor
2 (1) -1	vikingerik	2863	27 (29)	39 (38)	40	#1 on MB, #1 on Taxi, #1 on AFM, #1 on T2
3 (8) +5	Toolinit	2054	8 (5)	40 (29)	44	#1 on Fl2000, #2 on HD3E, #3 on HighSpeed, #6 on DrDude
4 (3) -1	Tarek Oberdieck	1770	19 (18)	20 (19)	21	#1 on BH, #1 on FH, #1 on Genie, #1 on DrDude
5 (9) +4	Sir_Moovenstein	1743	13 (13)	28 (25)	35	#1 on CV, #2 on AFM, #3 on ElDor, #4 on Genie
6 (7) +1	maestroreese	1674	4 (3)	34 (32)	43	#3 on DrDude, #9 on Genie, #9 on SpSh, #10 on CP
7 (4) -3	invitro	1624	3 (3)	31 (32)	43	#4 on FH, #9 on AFM, #10 on MM, #11 on BH
8 (6) -2	Space Tractor	1624	6 (7)	27 (29)	41	#3 on Pin*Bot, #5 on SpSh, #5 on FT, #5 on BRose
9 (5) -4	TOUGE_KILLER	1613	9 (9)	29 (30)	38	#2 on DrDude, #3 on Firep, #3 on WWind, #4 on CBW
10 (14) +4	dylan_h	1515	5 (4)	27 (27)	43	#3 on Co1812, #4 on DrDude, #7 on TeedOff, #7 on FT
Rank (previous) Change	Username	Points	# of Top 10 Scores (previous)	# of Top 100 Scores (previous)	# of Top 500 Scores	Best Scores
11 (10) -1	FALCO12-QUEBEC	1476	9 (11)	26 (26)	30	#1 on EatPM, #3 on RBioN, #4 on Centaur, #6 on CV
12 (11) -1	pinballchris	1469	3 (3)	28 (28)	42	#6 on ElDor, #7 on Co1812, #7 on CBW, #11 on NGG
13 (13) 0	Crusty Booger	1449	3 (4)	28 (28)	44	#5 on TeedOff, #8 on SpSh, #8 on Co1812, #11 on Victory
14 (12) -2	atoppachar	1441	8 (8)	23 (23)	38	#2 on FT, #4 on Taxi, #4 on GoinNuts, #4 on BRose
15 (15) 0	TRLBrony	1273	3 (2)	25 (25)	37	#6 on Co1812, #10 on Fl2000, #10 on HighSpeed, #15 on CBW
16 (16) 0	geohoo	1227	2 (4)	24 (23)	38	#6 on WWind, #9 on CC, #11 on FH, #12 on BK2K
17 (21) +4	ChristofferB	1200	4 (2)	23 (19)	38	#9 on GoinNuts, #9 on WD, #9 on HighSpeed, #10 on Victory
18 (17) -1	HUFMUF	1173	7 (7)	17 (17)	26	#1 on HD3E, #1 on TeedOff, #1 on Co1812, #2 on TZ
19 (19) 0	pinballwiz45b	1132	1 (4)	20 (17)	36	#9 on Victory, #11 on Co1812, #11 on FT, #13 on Pin*Bot
20 (23) +3	Just this guy	1069	5 (4)	20 (17)	24	#3 on WD, #4 on HighSpeed, #6 on BRose, #9 on HD3E
Rank (previous) Change	Username	Points	# of Top 10 Scores (previous)	# of Top 100 Scores (previous)	# of Top 500 Scores	Best Scores
21 (28) +7	Jazza	1042	1 (1)	19 (13)	40	#9 on Firep, #20 on CentPark, #22 on CBW, #28 on TeedOff
22 (18) -4	www.vpforums.org	1031	8 (10)	16 (16)	16	#1 on STTNG, #4 on AFM, #5 on BH, #5 on Genie
23 (133) +110	Fromduc	1006	0 (0)	20 (7)	31	#14 on SpSh, #14 on CBW, #17 on WWater, #17 on Pin*Bot
24 (27) +3	Tripple I	983	0 (0)	18 (16)	37	#16 on WWind, #17 on SpSh, #21 on CC, #21 on BRose
25 (20) -5	Nightmac21	962	0 (0)	18 (19)	36	#12 on TeedOff, #20 on CV, #24 on HD3E, #26 on CBW
26 (24) -2	MT	954	0 (0)	18 (18)	40	#14 on TeedOff, #20 on Firep, #23 on WWater, #26 on RBioN
27 (22) -5	DeeEff	950	0 (0)	14 (14)	39	#15 on WWind, #17 on Centaur, #21 on TeedOff, #22 on FH
28 (26) -2	CheesyBeefy88	922	1 (1)	18 (18)	34	#8 on BigShot, #15 on SpSh, #18 on CP, #18 on GoinNuts
29 (25) -4	PieCES	913	1 (1)	17 (18)	31	#10 on Firep, #12 on BH, #12 on DrDude, #32 on CV
30 (33) +3	ER777	871	1 (1)	15 (11)	39	#5 on ElDor, #16 on CP, #19 on Genie, #25 on HH
Rank (previous) Change	Username	Points	# of Top 10 Scores (previous)	# of Top 100 Scores (previous)	# of Top 500 Scores	Best Scores
31 (29) -2	vpalmer	871	2 (4)	16 (16)	26	#7 on SpSh, #7 on Centaur, #11 on Genie, #12 on FT
32 (30) -2	Howling_Goro	863	8 (8)	11 (11)	20	#2 on STTNG, #2 on Centaur, #2 on T2, #5 on BoP
33 (43) +10	smbaydp	863	2 (2)	12 (9)	40	#9 on Fl2000, #10 on ElDor, #20 on BigShot, #43 on Taxi
34 (32) -2	slam23	831	0 (0)	15 (14)	35	#19 on ElDor, #33 on FT, #34 on Victory, #37 on MM
35 (31) -4	chk	827	1 (1)	12 (14)	32	#7 on BRose, #12 on Fl2000, #15 on FH, #15 on BK2K
36 (34) -2	victorstulemeije	794	0 (0)	15 (17)	34	#12 on HD3E, #25 on TeedOff, #36 on Victory, #48 on CV
37 (92) +55	maezun	775	1 (2)	12 (8)	26	#8 on Pin*Bot, #11 on RBioN, #11 on ElDor, #17 on BRose
38 (35) -3	rockman_x_2002	757	2 (2)	12 (12)	31	#6 on GoinNuts, #6 on HH, #17 on Fl2000, #28 on Firep
39 (40) +1	Nemesis041	750	1 (1)	16 (15)	25	#8 on BK2K, #11 on WWater, #14 on HighSpeed, #26 on SS
40 (37) -3	gooche77	749	1 (1)	14 (13)	29	#7 on Genie, #24 on Firep, #29 on ElDor, #29 on FT
Rank (previous) Change	Username	Points	# of Top 10 Scores (previous)	# of Top 100 Scores (previous)	# of Top 500 Scores	Best Scores
41 (53) +12	ppeltola768	743	1 (1)	13 (10)	32	#9 on NGG, #14 on Victory, #28 on BK2K, #34 on HighSpeed
42 (36) -6	djchrille	740	3 (3)	13 (13)	23	#5 on T2, #7 on Victory, #10 on HD3E, #11 on BigShot
43 (50) +7	efbelo	724	1 (2)	10 (10)	30	#10 on CBW, #11 on MB, #11 on T2, #15 on CentPark
44 (38) -6	Caps_Lock420	712	1 (1)	13 (13)	26	#4 on ElDor, #20 on WWind, #28 on BigShot, #33 on Centaur
45 (41) -4	moldygoat	709	1 (1)	10 (12)	26	#6 on CP, #12 on WWater, #14 on BoP, #14 on WWind
46 (39) -7	Jake Alt	684	3 (3)	10 (10)	30	#3 on BRose, #4 on WD, #7 on BK2K, #22 on CV
47 (51) +4	Blah	683	0 (0)	10 (11)	35	#21 on DrDude, #29 on Fl2000, #30 on CC, #41 on WWind
48 (66) +18	phoerber	683	7 (6)	10 (8)	10	#2 on BK2K, #2 on HighSpeed, #3 on EatPM, #5 on CftBL
49 (42) -7	tinder	665	0 (0)	9 (10)	41	#13 on TeedOff, #40 on HD3E, #49 on AFM, #55 on SS
50 (45) -5	LordEngi	665	1 (1)	12 (12)	27	#10 on HH, #13 on CBW, #19 on GoinNuts, #26 on Firep
Rank (previous) Change	Username	Points	# of Top 10 Scores (previous)	# of Top 100 Scores (previous)	# of Top 500 Scores	Best Scores
51 (44) -7	torstenkuchenbae	663	0 (0)	13 (14)	30	#23 on DrDude, #24 on WWater, #37 on Fl2000, #42 on HH
52 (46) -6	ostrich2	654	4 (4)	13 (13)	17	#2 on CC, #8 on MB, #8 on HD3E, #8 on NGG
53 (180) +127	Benj341276	654	2 (1)	10 (5)	29	#5 on CC, #6 on MB, #28 on WWater, #32 on Fl2000
54 (48) -6	Chopin6868	653	2 (2)	12 (13)	23	#2 on WWind, #10 on Pin*Bot, #15 on HH, #19 on BoP
55 (52) -3	xNiCeGuYx	652	0 (0)	13 (12)	31	#15 on WD, #19 on Centaur, #20 on BK2K, #24 on GoinNuts
56 (47) -9	ParallaxScroll	644	0 (0)	12 (12)	20	#17 on MM, #18 on CV, #19 on Gorgar, #24 on BigShot
57 (49) -8	benika7	635	4 (4)	11 (12)	12	#3 on CV, #3 on MB, #7 on CC, #9 on TZ
58 (56) -2	Crazy Newt	635	0 (0)	13 (13)	28	#29 on HighSpeed, #30 on BRose, #32 on AFM, #42 on Pin*Bot
59 (57) -2	Fungi	634	0 (0)	11 (11)	29	#15 on Firep, #16 on HighSpeed, #50 on SpSh, #50 on CBW
60 (55) -5	Zoop25	624	0 (0)	9 (8)	31	#14 on Co1812, #28 on CP, #28 on GoinNuts, #39 on Genie
Rank (previous) Change	Username	Points	# of Top 10 Scores (previous)	# of Top 100 Scores (previous)	# of Top 500 Scores	Best Scores
61 (69) +8	RagingBull888	623	1 (1)	9 (6)	32	#2 on WD, #42 on HighSpeed, #69 on WWind, #70 on CentPark
62 (4933) +4871	Captain_Cockup	617	2 (0)	12 (0)	16	#6 on FH, #7 on RBioN, #11 on STTNG, #18 on BH
63 (59) -4	Vic Golf Chad	612	0 (0)	8 (10)	31	#14 on BK2K, #15 on TeedOff, #61 on Pin*Bot, #64 on SS
64 (54) -10	nikostos44	601	1 (1)	13 (13)	25	#3 on GoinNuts, #21 on Victory, #39 on SpSh, #47 on BoP
65 (62) -3	Mark_Miwurdz	582	5 (5)	9 (9)	13	#3 on BoP, #5 on RBioN, #7 on T2, #8 on FH
66 (76) +10	ryan	580	1 (1)	8 (7)	30	#8 on Victory, #35 on ElDor, #49 on TeedOff, #52 on Fl2000
67 (60) -7	kosacks	578	0 (0)	6 (6)	33	#25 on FT, #36 on TeedOff, #49 on Gorgar, #63 on Firep
68 (58) -10	Lordbobby	573	0 (0)	9 (10)	21	#12 on Co1812, #20 on SpSh, #25 on GoinNuts, #26 on Genie
69 (68) -1	kan	567	0 (0)	10 (10)	25	#25 on Co1812, #25 on HighSpeed, #30 on Fl2000, #30 on BK2K
70 (99) +29	pinbotwms	560	1 (1)	8 (5)	28	#9 on BRose, #20 on Fl2000, #33 on HighSpeed, #46 on Victory
Rank (previous) Change	Username	Points	# of Top 10 Scores (previous)	# of Top 100 Scores (previous)	# of Top 500 Scores	Best Scores
71 (61) -10	tuomas	555	1 (1)	10 (11)	17	#6 on AFM, #22 on SpSh, #24 on BH, #25 on Firep
72 (84) +12	mjohns	548	0 (0)	9 (8)	26	#38 on Co1812, #44 on HighSpeed, #45 on BRose, #51 on FT
73 (63) -10	RIZ420_420_1	547	1 (2)	11 (11)	16	#5 on WWater, #11 on Pin*Bot, #21 on CP, #26 on CC
74 (87) +13	krazysteve1958	545	2 (2)	6 (6)	22	#3 on FH, #5 on CBW, #52 on ElDor, #53 on HighSpeed
75 (77) +2	guitarguru80	544	0 (0)	10 (10)	22	#22 on Gorgar, #26 on HH, #34 on SpSh, #40 on Co1812
76 (64) -12	Roy Wils	542	0 (0)	11 (11)	16	#13 on CV, #16 on FH, #19 on HD3E, #27 on AFM
77 (65) -12	Phreaker47	542	5 (5)	6 (7)	15	#3 on Fl2000, #3 on HH, #5 on Taxi, #6 on BoP
78 (72) -6	PinHead970	533	4 (4)	8 (8)	10	#1 on WWind, #5 on STTNG, #5 on Co1812, #7 on TZ
79 (78) -1	siemas12345	530	2 (2)	6 (7)	19	#2 on SS, #9 on T2, #14 on HH, #42 on CC
80 (1377) +1297	Beatnik-Filmstar	528	1 (0)	9 (0)	24	#10 on Co1812, #33 on HH, #38 on FT, #41 on FH
Rank (previous) Change	Username	Points	# of Top 10 Scores (previous)	# of Top 100 Scores (previous)	# of Top 500 Scores	Best Scores
81 (71) -10	LUK	526	0 (0)	8 (8)	27	#13 on ElDor, #43 on CBW, #68 on Victory, #76 on BRose
82 (70) -12	kovalev1985	524	2 (2)	8 (9)	18	#2 on CV, #8 on Taxi, #19 on RBioN, #26 on Centaur
83 (74) -9	rehtroboi	522	0 (0)	6 (7)	29	#17 on Co1812, #48 on Victory, #49 on SpSh, #75 on FT
84 (75) -9	RZR	518	0 (1)	10 (10)	25	#11 on BK2K, #32 on WD, #42 on TZ, #56 on DrDude
85 (79) -6	carkid93	516	0 (0)	9 (9)	24	#16 on HH, #50 on GoinNuts, #52 on CP, #55 on WD
86 (67) -19	Inspector42	516	0 (0)	6 (6)	30	#30 on ElDor, #32 on GoinNuts, #52 on Genie, #64 on TZ
87 (371) +284	NeonZwei	511	4 (1)	7 (4)	9	#1 on BRose, #3 on AFM, #3 on FT, #6 on ToM
88 (82) -6	sneeking	507	2 (2)	9 (10)	17	#9 on Centaur, #10 on SpSh, #18 on Pin*Bot, #28 on Genie
89 (88) -1	Stonehedge16	501	0 (0)	7 (8)	27	#13 on BigShot, #19 on Pin*Bot, #50 on CentPark, #51 on GoinNuts
90 (73) -17	PrimeroIncognito	499	0 (1)	10 (10)	24	#11 on CP, #21 on Genie, #42 on GoinNuts, #58 on HD3E
Rank (previous) Change	Username	Points	# of Top 10 Scores (previous)	# of Top 100 Scores (previous)	# of Top 500 Scores	Best Scores
91 (89) -2	switch3flip	498	1 (1)	8 (8)	21	#6 on Victory, #21 on FT, #31 on WWind, #37 on CP
92 (80) -12	olivierarmy	494	0 (0)	12 (12)	14	#17 on MB, #18 on CC, #20 on CP, #34 on BRose
93 (81) -12	nonojin	491	0 (0)	6 (6)	27	#22 on HD3E, #49 on Taxi, #57 on DrDude, #58 on Co1812
94 (85) -9	tiotbellot	490	0 (0)	8 (8)	19	#11 on TeedOff, #18 on HH, #22 on T2, #23 on WWind
95 (138) +43	kay989	485	0 (0)	7 (4)	25	#41 on BRose, #43 on AFM, #43 on FT, #51 on DrDude
96 (83) -13	clawhammer	483	0 (0)	7 (7)	28	#20 on STTNG, #34 on WWind, #35 on RBioN, #38 on Centaur
97 (103) +6	aspiratatator	483	0 (0)	7 (7)	28	#20 on CBW, #30 on EatPM, #50 on Fl2000, #54 on HighSpeed
98 (110) +12	9u1d0	483	0 (0)	5 (3)	29	#13 on HH, #21 on BK2K, #26 on HighSpeed, #66 on TZ
99 (102) +3	Badasstronaut	479	0 (0)	6 (6)	29	#33 on GoinNuts, #38 on WWind, #40 on ElDor, #51 on Victory
100 (95) -5	nowclickthis	471	0 (0)	6 (5)	26	#26 on TeedOff, #34 on HH, #45 on Co1812, #64 on HighSpeed

 

[invitro]

Why do you require this? Is there a mathematical basis, or just doing it that way because you have been? All sorts of functions are easier to work out with a baseline of 0 rather than 1. The difference between 0 or 1 point at the bottom of the board won't mean anything for the total rankings and makes it mathematically cleaner.

Good question. I did it so I could compare it with the linear function. And I made it 1 rather than 0 to leave room for fractional points for ranks #101-500. I don't know if it was a good thing to do.

But anyway, that doesn't really matter. Yeah, what shape to make the curve is a bigger question. I think the log method is naturally emergent from mathematics, while the other non-linear methods are artificially constructed to fit what you/we already decided we wanted the curve to look like. The log method needs no magic-constant besides how deep you want to rank.

I don't think the log function is any more natural than the exp one. Actually I think it's less natural, since it goes to -infinity as rank goes to infinity, while the exp function goes to zero, which seems more correct.

I think we (people here) agree that a #1 rank should be worth more than two #50 ranks. I think the rest might just be a matter of taste?

Hey, how about we use the PAPA ranking method? http://www.ifpapinball.com/ranking-info We can treat each table as a tournament of equal value, skipping that whole complicated section about Tournament Value Adjustment. Then the ranking uses the sum of a linear method through the entire population plus a power method for the top 32 players. Interested in that at all?

I am interested. I don't understand it from a skim, but will look more later.

 

[Mark W**a]

There's been a mistake with my ranking.

I have the following high scores:

BR: 1st
FT: 3rd
AFM: 3rd
ToM: 7th (listed as 6th)
MM: 17th (not tallied)
CV: 35th (not tallied)
Twilight Zone: 50th (tallied)

My total points are being listed at 437, so you missed both my MM and CV scores.

Also, please stick with the original formula.

 

[invitro]

I think you guys are overcomplicating this. Just give out some bonus points for the top few places, maybe up to top 10, then the liner progression for the rest of the top 100, and half a point for whatever honorable mention slots you want to include up to 200 or whatever.

The number one spot should be worth significant bonus points though, then dropping off for second and third and disappearing by either fifth or tenth. Like 200, 150, 125 for the top three spots respectively would make sense. A number one spot should be worth a lot more than something from the teens or worse, that's the key point here.

Well, I suppose this would work too. I have a hard time feeling certain of much beyond that the fully-linear method is no good. I suppose I am enjoying trying to find the "right" method, and I also feel an obligation to get it right, such that I am ok with it and you folks are ok with it too. Maybe I should make a poll asking for the best method? (Just! Kidding!)

I see an advantage to bonus points in that a player can easily know how much a ranking is worth, possibly/probably more easily than plugging it into an exp or log function.

My current plans are to (1) look at the IFPA method, (2) choose that, the log function, the exp function, or your method, or a combination of those, (3) to finally curtail the number of tables that give points at 23 per player, and (4) write up a full explanation and start a new thread with that, the current rankings, and a history of players' rankings back to February (because history is good). Which I'll do soon, but not right now. Please bear with me hashing this issue to death , and I do understand any frustration.

 

[vikingerik]

I'm not exactly understanding the formula here or what it's meant to achieve, but I see myself going from 30th to 87th... uhh, behind some guy whose best scores are 50th something on tables.

I think something went wrong with the math. The log method is very heavily weighted to the top of the scale (one #1 is worth two #10) so yeah that doesn't sound right. #30 is worth only 23 points and #50 only 15, so I can't see how that could add up to over 500 total for Inspector42.

I still think the log method is more mathematically natural, because it doesn't have any magic constant to fit some artificial predetermined range. For negative results, just drop them. But yeah it is very top-heavy so we could use something else. The PAPA method does look promising, about exactly what we want, mostly linear plus a power distribution for the top 32 which seems just about right to peg the caliber of where exceptional players start.

 

[Mark W**a]

Well, I suppose this would work too. I have a hard time feeling certain of much beyond that the fully-linear method is no good. I suppose I am enjoying trying to find the "right" method, and I also feel an obligation to get it right, such that I am ok with it and you folks are ok with it too. Maybe I should make a poll asking for the best method? (Just! Kidding!)

I see an advantage to bonus points in that a player can easily know how much a ranking is worth, possibly/probably more easily than plugging it into an exp or log function.

My current plans are to (1) look at the IFPA method, (2) choose that, the log function, the exp function, or your method, or a combination of those, (3) to finally curtail the number of tables that give points at 23 per player, and (4) write up a full explanation and start a new thread with that, the current rankings, and a history of players' rankings back to February (because history is good). Which I'll do soon, but not right now. Please bear with me hashing this issue to death , and I do understand any frustration.

It's your decision and choice how you want to do it.

All I know is, I went from 30th (well 24th once corrected, see above post) to like 87th (probably around 77th once corrected) which is fine however I don't feel it's fair that I'm behind a guy whose "best scores" are in the 30th-60th range. One of his scores listed, Twilight Zone at 64th, is a table I happen to be 52nd on. But I don't even consider that one of my top scores at all. Even my 35th on CV is shakey. How can someone like that be ranked above someone with 4 top 10 scores? Doesn't seem fair to me, seems like it must favor those with a ton of tables, which I thought we were trying to get away from?

I think something went wrong with the math. The log method is very heavily weighted to the top of the scale (one #1 is worth two #10) so yeah that doesn't sound right. #30 is worth only 23 points and #50 only 15, so I can't see how that could add up to over 500 total for Inspector42.

I still think the log method is more mathematically natural, because it doesn't have any magic constant to fit some artificial predetermined range. For negative results, just drop them. But yeah it is very top-heavy so we could use something else. The PAPA method does look promising, about exactly what we want, mostly linear plus a power distribution for the top 32 which seems just about right to peg the caliber of where exceptional players start.

Ok. I figured.

These are my stats ivitro in case you missed them:
BR: 1st
FT: 3rd
AFM: 3rd
ToM: 7th
MM: 17th
CV: 35th
Twilight Zone: 50th

 

[invitro]

There's been a mistake with my ranking.

I can't find an error. Here are the points the exp formula gives you per table (before hackers are skipped):
BRose: #1 -> 100.0
AFM: #3 -> 91.1
FT: #3 -> 91.1
ToM: #6 -> 79.3
MM: #18 -> 45.4
CV: #35 -> 20.6
TZ: #51 -> 9.8
STTNG: #108 -> 0.7
BoP: #164 -> 0.1
total points: 438.0

(The table says 437 because that is after hackers are skipped, and also the integer in the table is truncated rather than rounded.)

Thanks for your feedback on the formula.

 

[invitro]

I think something went wrong with the math. The log method is very heavily weighted to the top of the scale (one #1 is worth two #10) so yeah that doesn't sound right. #30 is worth only 23 points and #50 only 15, so I can't see how that could add up to over 500 total for Inspector42.

The log method in the table uses POINTS = 100 * (1 - log(RANK)/log(500)), instead of having the log(104.76) from the earlier table. This pegs #500 to 0 points; the other pegs #100 to 1 point. Inspector42 has lots of scores in the #101-500 range, and #154 is worth 19 points with the log(500) version, so that's how he racks up over 500 total.

I used log(500) for the one I posted because I thought (maybe incorrectly) that's the one you preferred. %-) Inspector42 is ranked #286 when using POINTS = max(0, 100 * (1 - log(RANK)/log(104.76))), and #348 with the exp function.

 

[invitro]

How can someone like that be ranked above someone with 4 top 10 scores? Doesn't seem fair to me, seems like it must favor those with a ton of tables, which I thought we were trying to get away from?

I haven't implemented anything to limit the # of tables a player gets points for, yet; I will try to look at that and the IFPA system tonight or this afternoon.