Leaderboard Change Proposals - Discussion

[Weehawk]

I had planned to allow a little more time for discussion before formulating specific proposals, but since the original thread was fubarred I guess this would be a convenient time to start molding some likely candidates for implementation:

Proposition 1:

Do nothing.

One would vote for this if one is in favor of not making any change to the current system of awarding leaderboard points.

Proposition 2:

The old system.

Awarding points only for the top 3 places. 10, 3 , and 1 points respectively.

Proposition 3:

Using the current formula based on percentage of the top score scaled down 15% for each place below first (first getting 100 points). For those comfortable with formuale it would look like this:

pts = 100*(your_score/top_score)*.85^(n-1) where n is the place

but cutting off after 3rd place so that anything below 3rd gets 0

Proposition 4:

Current formula (see Proposition 3) but cutting off after 5th place so that anything after 5th place gets 0

Proposition 5:

Places get percentage of score scaled down by a subtracted 15% for each place, that is:

1st place gets 100 pts
2nd place gets (percentage of first place score) * .85
3rd place gets (percentage of first place score) * .70
4th place gets (percentage of first place score) * .55
5th place gets (percentage of first place score) * .40
6th place gets (percentage of first place score) * .25
7th place gets (percentage of first place score) * .10

anything below 7th place gets 0

Proposition 6

First five places get 10,7,5,3,1 pts respectively

Personally I would be leaning toward Proposition 3 or 4 myself as it stands.

Let us now discuss whether any of these proposals can be made more effective, whether any of the above would be better omitted, or whether there are other candidates left out which might have merit.

And let us pray that this thread doesn't get screwed up.

[destructor]

Position 6:
Totally percentage.

1st=100
2nd, 3rd and ...=percentage to 1st

[LN2]

Perhaps left off the list are:

instead of having the 10-3-1 for just the top 3, perhaps have 10-7-5-3-1 for the top 5 places instead.

Using proposition 3 formula, but counting that for all places even beyond the top 3 or beyond top 5 as long as the person has achieved at least 50% of the top score(or some other percentage maybe). This has a flaw though for games with an ending and max score....like Pole Position, Track-n-Field etc. where it's trivial to reach 50% of the top score.

Overall, I would be leaning more toward the 10-7-5-3-1 or proposition 5(which I had given reduced that big table to a formula)

Perhaps the discussion can aim at this point on whether you think points awarded should vary versus score relative to the top score. If so then that eliminates a couple of these choices.

I think the points should vary versus score relative to the top score. Absolute place points doesn't consider any of that.

Let's also not forget what started all of this discussion. It initially was about how to change the leaderboard point calculation so those that generally just submit a ton of scores playing each rom set only a couple times and submitting whatever score they get(whether good or poor score), what we termed ABC submitters/uploaders previously, no longer will be able to climb the leaderboard just from doing that.

This should be the #1 goal of any change.

[LordGaz]

Quote:
Having LB points decay over time is a non starter in my opinion. Just because a score is old in no way devalues it.

I think so too.

Just to squeeze in my last 2 penneth on this one, you all seem to have missed my whole point. I wasn't suggesting an old score is in any way worth less than one scored recently, I was merely suggesting another method for a leaderboard system. To reiterate, scaling down leaderboard points by age of recording will produce a leaderboard with a ranking of who are the best players now. If you make a few incredible uploads then disappear from MARP you will not stay high on the leaderboard you will have to continue to stay active in order to achieve that. What is wrong with that?

Anyway that said, I think the current system is ok so current I go for Weehawk's system 1 but I reserve the right to change my view. The only other real way to weed out the 'worthless' inps is to have an independent adjudicator/s to judge all inps by merit and award points accordingly.

Cheers all,
Gaz

[Weehawk]

Bleh.

Gaz's post didn't even update the "Last Post" link.

I could save the text and delete the post, letting him resubmit, but when I delete the post, the posts-to-thread counter gets decremented, even if that post hadn't incremented it in the first place

You could wind up with the thread thinking there are a negative number of posts to it.

[Weehawk]

Position 6:
Totally percentage.

1st=100
2nd, 3rd and ...=percentage to 1st

With no cutoff????

[Mr. Kelly R. Flewin]

Proposition 1:

Do nothing.

One would vote for this if one is in favor of not making any change to the current system of awarding leaderboard points.

Proposition 5:

Places get percentage of score scaled down by a subtracted 15% for each place, that is:

1st place gets 100 pts
2nd place gets (percentage of first place score) * .85
3rd place gets (percentage of first place score) * .70
4th place gets (percentage of first place score) * .55
5th place gets (percentage of first place score) * .40
6th place gets (percentage of first place score) * .25
7th place gets (percentage of first place score) * .10

anything below 7th place gets 0

I'm torn... the first option is what I feel works best, mind you we just change peoples option to be on the leaderboard if they're an Alphabet Uploader.

But then again, the 5th option doesn't sound too shabby but I'd need to see some examples to get a better idea of how totals would look.


Kelly

[destructor]

Position 6:
Totally percentage.

1st=100
2nd, 3rd and ...=percentage to 1st

With no cutoff????

Yes. No cutoff.

[Weehawk]

Position 6:
Totally percentage.

1st=100
2nd, 3rd and ...=percentage to 1st

With no cutoff????

Yes. No cutoff.

Wow.

That would be pretty much the opposite of what everyone else seems to be wanting to accomplish with a change, in that it would give a lot more leaderboard points to members submitting 17th place inps that are of no particular interest to the community.

I'm going to have to see some more support for this concept before I consider it a viable candidate.

[LordGaz]

Position 6:
Totally percentage.

1st=100
2nd, 3rd and ...=percentage to 1st

Fine for tournaments like the Deca but at MARP having some extra incentive to go for higher positions is a good idea IMO.

[LN2]

But then again, the 5th option doesn't sound too shabby but I'd need to see some examples to get a better idea of how totals would look.

Kelly

Zwaxy posted LB points for 1942 using the current system and prop 5 as an example in that other thread.

In short, it doesn't affect the top 3 much at all versus the current system. It makes the 4th and 5th places only slightly lower...6th and 7th places quite a bit lower...and 8th place and beyond all zeroed.

o reiterate, scaling down leaderboard points by age of recording will produce a leaderboard with a ranking of who are the best players now

...cuz I and I think most believe the LB should reflect the best players...not the best active ones.

Those ABC uploaders are the "best active ones". I think the overall net effect would be to make them even higher ranked than they currently are by constantly submitting scores...even if good for only a few points here and there. They would still earn more points than many masterful high scores submitted 2-3+ years ago.

Down the road most that have established lots of records here likely won't be as active as they have been in the past. Perhaps they are playing some, but since they already have first place scores, aren't going to be submitting any new scores for those games most likely.

It also almost forces the best players to find other games they might not like and just play them for spite just to get some LB points to try and stay near the top. At some point it would become impossible for them unless they went and beat all of their previous personal high scores. That seems like a lot of effort to just stay at/near the top.

...in a sense making the best players closer to ABC uploaders also by submitting scores for various other games...or replacing their current ones with possibly lesser scores.

If you make a few incredible uploads then disappear from MARP you will not stay high on the leaderboard you will have to continue to stay active in order to achieve that.

if you only make a few uploads you won't be very high on the leaderboard anyway where anyone will care if you are in 75th or 325th. If you make 100 top uploads, then never submit again, odds still are you slowly will drop down the leaderboard and also have some scores beaten so drop points also.

roncli is a great example of this. 1.5 years ago he was in 9th or 10th place with around 11,000 LB points. Now he is 26th with 8,200 LB points. Even if he still had all of his 11,000 points he would be ranked around 19th instead of 9th or 10th his 11,000 gave him 1.5 years ago.

Therefore, from his inactivity and some beatable scores, his LB points have "decayed" without needing any special decay formula etc.

If someone has all scores truly untouchable, then their LB points should be untouchable also. Even with that, they slowly will still drop down the leaderboard cuz other players are submitting more scores and increasing their LB points. Also with new rom sets that are added or made working in MAME that allow all active players to increase their LB points.

[Buttermaker]

Proposition 1

Do nothing.

One would vote for this if one is in favor of not making any change to the current system of awarding leaderboard points.

This is unacceptable and should not be an option. We're doing all this here to get rid of that system.


Proposition 2

The old system.

Awarding points only for the top 3 places. 10, 3 , and 1 points respectively.

I like this because it's like a podium you see in all kinds of competitions.

10-3-1 might not be the best way to distribute the points. Anybody got a better suggestion?


Proposition 3

Using the current formula based on percentage of the top score scaled down 15% for each place below first (first getting 100 points). For those comfortable with formuale it would look like this:

pts = 100*(your_score/top_score)*.85^(n-1) where n is the place

but cutting off after 3rd place so that anything below 3rd gets 0

Very good. This one or Proposition 2 is the way to go.


Proposition 4

Current formula (see Proposition 3) but cutting off after 5th place so that anything after 5th place gets 0

Not good. Too many chances for ABCers to get LB points.


Proposition 5

Places get percentage of score scaled down by a subtracted 15% for each place, that is:

1st place gets 100 pts
2nd place gets (percentage of first place score) * .85
3rd place gets (percentage of first place score) * .70
4th place gets (percentage of first place score) * .55
5th place gets (percentage of first place score) * .40
6th place gets (percentage of first place score) * .25
7th place gets (percentage of first place score) * .10

anything below 7th place gets 0

In time every game will have 7 decent scores for it, and the ABC uploaders will give up. Are we willing to wait for that?

No. Just think about the 100 (yes, 100) Atari clones added in a recent version. There are way too many games/clones in MAME for MARP to have 7 decent scores even for a fraction of all games. I also want those inps to not even get uploaded in the first place.

[zlk]

I am in favor of the 10-3-1 system. If you want to do well on the leaderboard, you will need to spend some time making a quality recording. If your score isn't in the top three but you still think you played well and want to show others the replay, you can then upload the recording for no points. I think this system would solve many problems.

[Weehawk]

Proposition 1...This is unacceptable and should not be an option

The only way to prove that it is unacceptable to the community is to include it as an option...and for people to vote for something else.

Proposition 2...10-3-1 might not be the best way to distribute the points

True, I feel that there is too much penalty for 2nd place there.

Also, I personally think that any system where a second place score which is 99% of the first place score gets the same leaderboard points as a second place score which is 1% of the first place score is flawed to such a degree that I would never vote for it.

There does, however appear to be support for it, so I am tentatively including it. It could be modified.

[LN2]

It's hard to believe the number of options that essentially state to give LB pts to the top 3...and screw everyone else.

That's a bit harsh just to keep ABC uploaders from getting points.
There aren't very many of those. I looked down the leaderboard today and only saw a handful who are ABC uploaders.

There are far more decent and quite good scores that are 4th, 5th, 6th versus those places being held by ABC uploaders.

Sure, there are lots of games where it doesn't take much to get 6th or better...so what? Eventually more will submit scores for the game so it will take a decent or quite good score to just make 6th place.

Look at that 1942 example Zwaxy had again. If you only award points to the top 3, then those that are also experts and completed the game will get ZERO.

This is similar for many of the golden era more popular games.

This potentially will discourage some participation by having the top 3 only count where more will figure they can't compete with the top 3 so won't bother.