This is my opinion as far as how the leaderboard should rank all the players. Each 1st place score on a given game should be equal to 100% and then each score there after would be a lower percentage based on how high the 1st place score was. Twin Galaxies utilizes this ranking system and has been for almost 20 years. It works the best and is very simple to follow. Yes it will change the leaderboard around a little, but the players that will benefit from this method will be the ones who have the most number of scores for games available to date. I gave another example below as I did back some time ago in another message I wrote. I think this is the easiest way of fairly ranking each player score based on the number of scores available for a given game. Rather then the 10,3,1 ranking system.

Regards, steve Krogman

Galaga: 5405930 me 100% 1630780 beejay 30.2% 625720 thunky 11.6% 149270 roto 2.8% 52420 richard kev stewart 1.1%

Donkey Kong Junior Bootleg:

485700 notman 100% 273300 kft 56.3% 118700 randood 24.4% 85500 me 17.6% 41300 oldie 8.5% 24300 ron 5.0% 12600 goldcrisp 2.6% 8400 slayer 1.7%

pangworld

2847830 me 100% 2700000 xtc 94.8% 2638080 renzo vignola 92.6% 1516570 angry 53.3% 481030 n rodgers 16.9% 461690 barry rodewald 16.2% 75770 sorceress 2.7% 57260 zwaxy 2.0%

-- stephen krogman (skrogman@concentric.net), June 12, 1999

Answers

I've been thinking about my previous post on no more than 15% for the 1st place bonus (or other score penalty).

I still didn't like any penalty - after all the scores are what they are. End of story. If someone gets 98% of your score why shouldn't they get 98 points. This is especially true of the bowling games. Why should someone get more points for the same score just because they were the first to upload a perfect 300 game?

My vote is as per TGK's above:

Top score = 100 points. All other scores = percentage of top score.

While it's true this will encourage quantity if someone puts in a crap score for a game, then someone else can take the points they earn down to size by simply uploading a new recording. I think the fact that every point extra you can get on games where you are currently 2nd will actually encourage you to play games that you may not be able to beat the top score on now, but at least you're getting rewarded for every improvement you make. I think competition will improve with this method.

If someone wants to spend the time uploading lots of recordings then they should be rewarded. We all have the same opportunity for uploading tons of recordings and who knows we may find games we actually like and can get better at.

Now, time to try doing some more battling with BIL - who seems to like taking away my Galaxian scores...... ;-))

BeeJay.

-- beejay (bjohnstone@cardinal.co.nz), June 12, 1999.

Simple, obvious, time tested, bleeping traditional standard scoring method ... no, of course we can't use that...

I propose we use Aqua's Aggravating Algorithm instead, where points are assigned based on the squares of the first ten numbers in the Fibbonacci Sequence (0,1,1,2,3,5,8,13,21,34) to assure the proper rewards for good play and quick uploading...

Galaga: 1. Steve Krogman 1156, 2. BeeJay 441, 3. Thunky 169, 4. Roto 64, 5. Richard Kev. Stewart 25

Comotion: 1. 4 Phil Lamat 1164, 2. 4 Lagavulin 441, 3. 4 German Krol 169

Donkey Kong Jr. Bootleg: 1. Notman 1156, 2. Kft 441, 3. Randood 169, 4. Steve Krogman 64, 5. Oldie 25, 6. Ron Corcoran 9, 7. Goldcrisp 4, 8. Slayer 1

Dominos: 1. 4 JoustGod 1164, 2. 4 German Krol 441, 3. 4 Angry 169

Pang (World Tour): 1. Steve Krogman 1156, 2. Xtc 441, 3. Renzo Vignola 169, 4. Angry 64, 5. N Rodgers 25, 6. Barry Rodewald 9, 7. Sorceress 4, 8. Zwaxy 1

9th place would also get 1 point, 10th would get 0. If MARP desires, we could even extend the sequence so 11th and below get negative scores...

Take a wild guess,

-- Aquatarkus (aquatarkus@digicron.com), June 12, 1999.

Ok, i really like the percentage scoring, but i have to add that maybe the TG system has "worked" for 20 years, but who says it couldn't have worked better if there was a little bit more incentive for beating someone's score.

I think right now i wouldn't mind with a plain percentage OR a percentage with a little benefit between the higher places. And you can't just disregaurd the rigorous battles for placement we've had with the 10 3 1 system, that totally boosts the quality of recordings. We may see less of those battles or we may not, but we know the 10 3 1 system incited battles.

The bowling thing and scores that tie don't need to adjust people because they had a 300 after someone else had 300, you just adjust the 2nd place score which ever score is less than 300, i.e. you make 2nd place the second place score not the second highest scoring person.

It is unfortunate that there are games with a maximum score that is easy to get like the 300's for bowling but i think that's something we have to live with.

-- Chad (churritz@cts.com), June 12, 1999.

Aqua, now you are getting pretty tough, we don't need anything exponential... I love the old 10-3-1, makes you strive to be the 1st, and win the extra 7pts. BTW, I will organize the tourney if nobody else wants to organize it.

-- Chris Parsley (cparsley1@hotmail.com), June 12, 1999.

Well seeing that we're in the topic of complex scoring systems... I'll offer mine.

(The integral of (2x+5)) * (5x+4)* sin(7x^2+12x+14y^3+12y+493) + 1,293,001 ---------------------------------------------------------------------- --------- 14,495 * cot(19x^3+12y^2+1)

all + 1 Billion if your name is Gameboy9 :)

x is the number of people you beat in a particular game. y is the number of percentage points you would earn in a game.

The name? The Most Ridiculous Looking Leaderboard You've Ever Seen. It's guaranteed to give me first place overall :)

......... OK... so that's not a good idea... I'm just having a little fun :) Thanks for reading :)

-- Gameboy9 (goldengameboy@geocities.com), June 12, 1999.

Ok, a more serious answer from me but an even less helpful one.

It doesn't make a smegging bit of difference what sort of point system we use if we don't adjust it for games it doesn't work on. Chris Parsley's Superstars proprosal struck such a nerve, because to me it seems like replacing one rather arbitrary scoring system with another. The key advantage the percentage scoring system has is assigning a nonzero value to all scores, present and future, as a method of gauging how good all the contestants are. Problem is, the percentage system is just as arbitrary. As he pointed out, the 1400+ games do not have similar enough scoring methods to rank properly... he just missed that no simple blanket system will.

The first, and should have been the easiest, category are junk games. There are two types of these. The first is non-playable games, of which MACH 3 is sadly the most glaring example. The second is non-competitive games, in which all players "achieve" the same score. I hadn't noticed these until listing the D games, I found Dominos right under Domino Man, with a breathtakingly high score of 4, which was somehow reached (after countless milliseconds of developing their skill to the exacting standards required) by no less than 3 players. If you weren't paying attention, ALL games of Dominos end 4-0.

The second category are like the bowling or trivia games that aren't exactly easy but with some work all good players can have perfects. The only real difference between scores in this sort of game is the time of filing, which is probably what the rank should be based on.

The third category are the games with "Ninth Key" levels, but no forced ending, where after a certain point the game doesn't get any harder and you can play until you give up. Mat Mania hits this point somewhere around Match 20, depending on the difficulty setting. After about 300,000 points all Mat Mania scores are demonstrations of stamina.

A fourth category are games with "Ninth Key" levels and a forced end. Pac Man, the game this reference comes from, has about 20 unique levels, then 200+ repetitions of "Ninth Key", then finally an end screen. The worst plausible finishing score, by a player who didn't know how to catch a single monster but caught both keys every time with their "Ninth Key" pattern is probably around 3 million points, or approximately 90% of a perfect game. Although a few players have come very close, no perfect games yet. Is 3,000,000 worth 90% of what a perfect score recieves?

The fifth category are games which have non-repeating levels and forced endings, such as 1942 or Black Tiger. These games could be judged either on the number of levels completed (all finish scores equal), or ranked by the amount of points the players can squeeze out of them which inevitably favors the most recent entrant.

The sixth category are games with a number of levels repeating in a loop. Scramble, Gyruss, Gladiator, and I suspect every game Namco ever made (at least from the early to mid 80s) fall into this group. Once all the levels of the hardest loop are mastered, these games are demonstrations of stamina. Perhaps a Track & Field "One Round Only" rule for these?

The seventh category are games with no known point of maximum difficulty, repetition, or forced end. There is a good chance the games that seem to belong in this category (like Super Punch Out or Domino Man) have not been completely mastered yet, but until then the percentage scoring works very well on these.

The eighth category are timed games, where the high scores are often quite low compared to other games. We know the best Depth Charge player is German Krol, but we don't really know just how amazing (or if) his score of 3,510 is, but it has to be much more than twice as hard then scoring 1,755. Many sports games fall into this category.

I've probably missed a few groups, but by now the idea should be clear. We either settle for a simple system, or we use our collective expertise to come up with appropriate rules and methods for each type of game seperately. For instance, category 7 games might have straight percentage scores (0 to 100) that go into the leaderboard, where category 5 games might send the levels_completed / end_level as the score (the #1 ranked game not neccesarily being a 100 if nobody on MARP has finished it), but category 8 games only score a top five (#1=100, #2=80, #3=60, #4=40, #5=20).

So, which answer was more annoying?

Aqua

-- Aquatarkus (aquatarkus@digicron.com), June 13, 1999.

Aqua, very good on the Pacman percentage on a perfect game. 3,000,000 is just about 90% (0.899928000576 to be exact :) ) of a perfect game in which Rick Fothergil of Canada missed it by 90 pts at the New Hampshire tournement last month.

Just thought I'd jump off the subject for a sec. :)

Regards, Steve Krogman

-- stephen krogman (skrogman@concentric.net), June 13, 1999.

Steve,

The Twin Galaxies percent system works best in a tournament situation. That is: when ALL the players play ALL the games. The original Coronation Day tourneys had only five games and all 30+ players played three times for a score no matter how good or bad. Move over to MARP and there's now 1400 games! Since it's unrealistic for everyone to submit this many game recordings, I think a penalty system would make up for this and stir up competition between the various place holders for each game. See my append in the leaderboard comments thread...

-- Pat (laffaye@ibm.net), June 13, 1999.