Page 1 of 1

Old ranking system

Posted: Wed Sep 05, 2018 7:40 am
by ansollena
I do not know about you ... but I regret the times when there was the old ranking that gave the points based on the gap from the first classified. And then being fifth or tenth, doing more or less points, it was funny. Today, if you can't enter the first three positions, it is useless to play. Even often it is useless to finish second or third, because the only thing that really matters for the classification is to finish first. What a pity!

Re: Old ranking system

Posted: Wed Sep 05, 2018 9:44 pm
by ***PL***
I agree. The current ranking system is more about quantity and not quality.

Re: Old ranking system

Posted: Thu Sep 06, 2018 1:51 pm
by Cicca
The actual system is actually the original one.....
I also remember the times when we decided (I was one of the promoters) to introduce that alternative leaderboard (don't remember if a specific name were given)....it lasted some time, with its pros and cons, like any other.....and then dismissed...don't know why, I had been away from MARP for a while ....

Re: Old ranking system

Posted: Fri Sep 07, 2018 3:55 am
by RAX
In case of modifications of the actual leaderboard, I would suggest a kind of grouping (parent+clones). If a parent has a million top score, it's awful that someone gets a 1st place with a 1K score on a new empty clone. (exceptions for clones with different scoresystems).
And some games has 20+ identical clones: one training=20 1st places

Re: Old ranking system

Posted: Tue Sep 18, 2018 7:23 am
by Novice
RAX wrote: Fri Sep 07, 2018 3:55 am If a parent has a million top score, it's awful that someone gets a 1st place with a 1K score on a new empty clone. (exceptions for clones with different scoresystems).
And some games has 20+ identical clones: one training=20 1st places
DEFCON=6
If you want ,then you may do so.
I am not interested in huge amout of junks as I send.

Why not try marvel land ,RAX?
it is not a easy game.(I think)

Re: Old ranking system

Posted: Tue Sep 18, 2018 9:03 am
by INNUENDO
Hello, guys.

In my opinion, to avoid these junkies uploads, we can work only with the original games. The best is: We don´t have to full the site with a lot of inps.

I vote YES for this. :)

Re: Old ranking system

Posted: Tue Sep 18, 2018 6:07 pm
by BeeJay
INNUENDO wrote: Tue Sep 18, 2018 9:03 am we can work only with the original games. The best is: We don´t have to full the site with a lot of inps.
The difficulty with this idea is what happens when the 'original' version of a given game changes with new releases of Mame. Also, sometimes the 'original' versions had bugs which were fixed in later versions and therefore the later version is the better version to use for high score competitions etc. Then, there are games that while notionally 'clones' of the 'original' have quite different game play to the 'original' version.

Where games truly are identical, or so close to identical it has no discernible difference in game play, it does seem somewhat pointless to have so many of them available for upload.

My vote would be to completely get rid of the current leaderboard as that is what encourages so many of these crap uploads in the first place. If they didn't 'climb' that leaderboard by submitting that crap on unplayed games, than they would have far less reason to want to submit that crap... although we'd still have the appalling behaviour currently displayed by 'saulonpb' where they seem to be determined to put up a recording for every single game regardless of the impact on their leaderboard standing. Personally I find that such a direct violation of rule 1e that they could/should have been banned by now to remove the likelihood of real uploads being missed when they're removed from the recent uploads by this flood of absolute crap from saulonpb.

Cheers,
BeeJay.

Re: Old ranking system

Posted: Wed Sep 19, 2018 2:06 pm
by Haze
INNUENDO wrote: Tue Sep 18, 2018 9:03 am Hello, guys.

In my opinion, to avoid these junkies uploads, we can work only with the original games. The best is: We don´t have to full the site with a lot of inps.

I vote YES for this. :)
who decides what's worthwhile and what isn't

MAMEdev don't make any attempt to claim that the parent version of a game is the best version (in 99% of cases the Japanese clone is the best version of any given game, but there are literally hundreds of interesting clone sets in MAME even beyond that)

Re: Old ranking system

Posted: Fri Sep 21, 2018 5:38 am
by RAX
Novice wrote: Tue Sep 18, 2018 7:23 amWhy not try marvel land ,RAX?
it is not a easy game.(I think)
Indeed I love it, but I was not so good in scoring. Some time needed to do things in a satisfying way. I have an infinite to-do list, and Marvel Land is quite high in it.

Re: Old ranking system

Posted: Fri Sep 21, 2018 2:04 pm
by ***PL***
I think this idea had been suggested before, but here is my thought on the subject...

MARP members could always pick the one version to count towards the leaderboard. The default would be the parent, but in the case where a clone is actually better or if the parent has been renamed (where the highest scores had been submitted), then only that clone would count. Again, for each game, only one ROMset counts for LB points and the others are completely ignored. All splits for the chosen ROMset would count as well.

Re: Old ranking system

Posted: Sat Sep 22, 2018 4:05 pm
by Haze
if it were my choice I'd just scrap the leaderboard in it's current form as it's too open to abuse by poor players who want to look good

instead do it more like a competition based system, have a set of say 30 romsets each month, count scores played on those games during that month only. it would encourage people to play things they might not otherwise play, and make all this spamming for leaderboard places go away.

I don't think a genuinely good player needs the ego-stroking of the leaderboard in it's current form anyway, if you're a good player, you know it, and everybody else knows it, because you've got inps worth watching on here.

Re: Old ranking system

Posted: Mon Sep 24, 2018 9:01 am
by INNUENDO
BeeJay wrote: Tue Sep 18, 2018 6:07 pm
INNUENDO wrote: Tue Sep 18, 2018 9:03 am we can work only with the original games. The best is: We don´t have to full the site with a lot of inps.
although we'd still have the appalling behaviour currently displayed by 'saulonpb' where they seem to be determined to put up a recording for every single game regardless of the impact on their leaderboard standing. Personally I find that such a direct violation of rule 1e that they could/should have been banned by now to remove the likelihood of real uploads being missed when they're removed from the recent uploads by this flood of absolute crap from saulonpb.

Cheers,
BeeJay.
I agree with you. And I vote the same like you, BJ. Thanks for answer me. :)

Re: Old ranking system

Posted: Sat Sep 29, 2018 10:37 am
by Kale
fwiw I don't think saulonpb is mass uploading for the leaderboard.
Instead he's probably just abusing the search system: by adding a description in his garbage inps with "first try" it's basically "tagging" games with an arbitrary "popularity Arcade ranking" schema.
Think about it: "first try" means more often than not being last in games, which translates to his position in a specific game being a sortof rating ... :roll:

As for limiting this kind of misconduct: you might add a function to the LB, giving an arbitrary malus after a specific threshold if being in last place and not zeroed.
For example, here's the current situation with galaga romset, cut down to show the disparity:

Code: Select all

galaga (Galaga (Namco rev. B)) #1st : 6800910 Stephen M. Krogman (100%)
galaga (Galaga (Namco rev. B)) #2nd : 5782510 Barra (72%)
galaga (Galaga (Namco rev. B)) #3rd : 4351080 gastrainga (46%)
galaga (Galaga (Namco rev. B)) #4th : 3899880 GSH (35%)
galaga (Galaga (Namco rev. B)) #5th : 2121890 DGO (16%)
[...]
galaga (Galaga (Namco rev. B)) #40th : 52420 richard kev stewart (0%)
galaga (Galaga (Namco rev. B)) #41st : 48420 wolfman24 (0%)
galaga (Galaga (Namco rev. B)) #42nd : 48000 GoldCrisp (0%)
galaga (Galaga (Namco rev. B)) #43rd : 0 PAC (0%)
galaga (Galaga (Namco rev. B)) #43rd : 0 The_Galaga_King (0%) 
GoldCrisp will get a x malus, while PAC and The_Galaga_King will get an y malus in their LB points.
We can work out how much x and y are worth, and make x = y or not. imho cheaters are worse but it's debatable.

You can apply a similar malus system for uploading a 1st place on a game where cross-comparing the scores will result in a last place in parent set. For limiting the "stage 1 clears" by a certain individual ... :?

Then again: statistical wise comparing apples with bananas will never be perfect, therefore a perfect LB will never ever happen ...