LEADERBOARD, What if...?

[RAX]

Hi MARPers

Many of us told that the board is often ruined by useless submissions, that hide good replays among Gambling&Shit.
Moreover the Leaderboard get worse, due to this flood of nobody_is_going_to_play_them inps.
And climbing the Leaderboard seems the only reason to submit them.
Ok, let's try to do something.


In my opinion gambling games are pure shit, let's cut'em all:
http://www.arcade-extreme.com/public/MA ... mbling.htm

I don't like table games like mahjong, card games, etc. Lets' filter them:
http://www.arcade-extreme.com/public/MA ... oTable.htm

In my opinion even multi-games are bad. If you dislike Touchmasters and MAMEcab in MAME, then Sim Sala Bim:
http://www.arcade-extreme.com/public/MA ... iGames.htm

Many new clones have RIDICULOUS fresh first places compared to parent's or other clones' top score. Shall we talk about 80 points at Cadash??? In my opinion if you don't reach at least 25% of the top-group-score your score is worthless to a leaderboard:
http://www.arcade-extreme.com/public/MA ... ercent.htm

Whoa, the effect of those filters is nice, let's try to apply them all together
http://www.arcade-extreme.com/public/MA ... ercent.htm
(it's my favourite )

Hey, let's take a look to the special only_filtered_scores_leaderboard.
http://www.arcade-extreme.com/public/MA ... ording.htm


Maybe we could try some other filter, take a look here.
http://www.arcade-extreme.com/public/MARP/


Filters are applied using these lists:
casinò http://www.progettoemma.net/indice.php? ... a=&genre=6
quiz http://www.progettoemma.net/indice.php? ... a=&genre=7
multigames http://www.progettoemma.net/indice.php? ... =&genre=10
table-games http://www.progettoemma.net/indice.php? ... =&genre=13
mini-games http://www.progettoemma.net/indice.php?page=3&genre=18


According to MARP policy, these categories do not count for leaderboard:
- */archived romsets submissions
-zeroed scores
-K3, K4 and tournaments submissions

There are some minor strange inconsistencies between board and official Leaderboard.
Example: Phil Lamat has 946 1st places in leaderboard, while having 1004 1st place submissions (59 archived): it should be 945.
BBH has 266 1st places in leaderboard, while having 268 1st place submissions (3 K4's inps): it should be 265.
Gastrainga has 29 1st places in leaderboard, while having 27 1st place submissions.
There are few of these "errors" around, but I coul have misunderstood something: at the moment I re-calculated the general Leaderboard before filtering it.

Unfair eliminations happen with these automatic filters, I know that.
Three wonders? Sorry, multigame, filtered
Mortal Kombat? Top group score is that 99'999'999 on fucking clone
And so on. Every single case should be discussed, It's not my purpose here. I really wish that something moves on in MARP Leaderboard calculation.

I will update these UnOfficial leaderboards monthly.
I can also implement other kind of filters, if you wish: tell me

[dan]

[gameboy9]

That's awesome. Well done. My personal filter that isn't included? No gambling and no clones.

I wonder... do you think you could summon up a leaderboard using the super-old leaderboard system? This means 10 points for 1st, 3 for 2nd, and 1 for 3rd, and 0 for everybody else.
I'd also wonder what the HARP system would wind up as... that's 20-10-6-4-2-1-1-1-1-1-...

... but both of those are just for my curiosity. No problem if you don't want to do either.

[sawys]

Excellent !

I love your leaderboards

[Frankie]

I'm pro banning Luck games, but I'm against banning Mahjong. Mahjong is skill I believe.

Maybe just give points for Parent rom, and then give points by how many players have submitted for the game, but with a max for top 10 players, so that those outside top 10 have something to fight for Meaning if a game have a submitted score from 3 players they'll get 3,2,1 points, and 10,9,8,7,6,5,4,3,2 & 1 points for games where there's 10 or more submissions. Or maybe 20,18,16,14,12,10,8,6,4 & 2 points.

Or continue as it is now, and ban Luck games and clones on the Leaderboard.

[mahlemiut]

I like the idea of having a customisable leaderboard, just have to figure some way of categorising the games, then poke Zwaxy into life to implement it.

[BBH]

In my opinion gambling games are pure shit, let's cut'em all:
http://www.arcade-extreme.com/public/MA ... mbling.htm

Look how far el wil falls under this leaderboard

[RAX]

Leaderboards updated
http://www.arcade-extreme.com/public/MARP/

@Gameboy9, here is yours http://www.arcade-extreme.com/public/MA ... Clones.htm

I've been a bit busy in these days, next week I will add your old leaderboard system request (10-3-1-0-0 & 20-10-6-4-2-1-1)

[gameboy9]

@Gameboy9, here is yours http://www.arcade-extreme.com/public/MA ... Clones.htm

I've been a bit busy in these days, next week I will add your old leaderboard system request (10-3-1-0-0 & 20-10-6-4-2-1-1)

That's fine by me. Thanks for looking into it!

[LN2]

I see lots of the pacman clones got eliminated with these filters...have to be careful here...many score differently or even have an early game end screen so the potential scores are not the same.

take a game like the pacman clone piranha.... my scores are discredited with this filter cuz it's a clone that scores less than 25% of the top of the parent. well duh...piranha is so different no way you will score as high...or even close.

I agree with parts of it...although the even older percentage points system still rocked IMHO.

also, there are games at MARP where special rules are setup for 1 of the ROM sets...but not the others which may lead to the top score for that 1 set being much less than the others. ie. 5 man vs marathon.

there are many examples of this where some clones are so different they play a lot differently and have lower scores because the game is harder.

like the overall concepts considered here...but have to make many exceptions for some clones not sure it is worth it.
think of games where 1 clone is trackball/analog and another is digital/keyboard. a couple score much differently due to using the 2 methods of control.

[RAX]

Your points are right, and I agree with you.
But we all are saying in the last years that there are some problems with gambling, clones, poor scores (IMHO with multigames too), and nothing is moving.
My system would need a white-list and a black-list, impossible for me alone to write them.

The no_clones and only_best_score filters *are* unfair, I dont' like them too much, but we can't have romset parent-only and romset with 23 clones into the same leaderboard

I would prefer that every romgroup (parent+clones) goes into a single scoreboard. When all games are identhical (even with some language difference) we can compare scores in a single list. Clones with strange scoresystems or easier or non-matchable should be archived.


Leaderboard updated
http://www.arcade-extreme.com/public/MARP/

[RAX]

Updated to 18 January, with new MAME0.141 lists

[***PL***]

RAX, this is very cool, thanks for tracking this!! It reminds me of MARP leaderboard conversations back in 2001-02. I remember both gb9 and cicca kept lots of funky stats, like average points per game and other neat tourney stuff. There were even stat links on the header page too! At the time, we had a scoring system where every INP was worth 0-100 points. There was a penalty of 15% for each place you fell behind. This was the formula actually used at MARP...

(your score / high score) * (.85)^(place-1)

There were many proposals to change it, but we became divided in that some really liked the existing system, and others wanted something new and different. We looked at counting only the parents (originals), but it was noted that sometimes for a game one of the clones should be the true parent. Occasionally, MAMDEV would have the wrong parent (and still do), so it was deemed too difficult to track which ROMset to count in the LB standings.

Some wanted even more emphasis on position, so they thought F1 racing scoring would be interesting. At the time, it went 10-6-4-3-2-1. Now I see it's 25-18-15-12-10-8-6-4-2-1 for the top 10 places according to Wikipedia. http://en.wikipedia.org/wiki/List_of_Fo ... ng_systems

3-2-1 was considered and rejected, and later I came to the conclusion that 6-3-2 would actually perform better because a win is worth double 2nd place and triple 3rd place. I believe Tennis and Golf majors use similar ratios for handing out prize money.

In the end, as MARP is democratic and listens to its players... by majority vote, we decided the current 1-1-1 (i.e. modified Olympic medals total) would be the best! At the time, there weren't all those crappy games, but of course that is another topic...

[Cicca]

RAX, this is very cool, thanks for tracking this!! It reminds me of MARP leaderboard conversations back in 2001-02. I remember both gb9 and cicca kept lots of funky stats, like average points per game and other neat tourney stuff. There were even stat links on the header page too! At the time, we had a scoring system where every INP was worth 0-100 points. [cut]

I'm really pleased someone still remembers the hard work I did in keeping statistics for the first 4 Tournaments, after a long period of discussions, proposals, and so on.....

I recently opened a topic on the matter, but it got no replies....
http://www.marpirc.net/viewtopic.php?f=2&t=14789