202 replies to this topic

[Charons Little Helper]

Kaldor, on 13 April 2014 - 08:13 AM, said:

No, they do not need to start them in the middle. Drop them down 500 points or so, so they really are in the starter ranks. If they are good enough, they will get to mid level and above quickly.

Whatever score they are given will eventually become the middle. It's how Elo works.

[Kaldor]

Charons Little Helper, on 14 April 2014 - 05:45 AM, said:

Whatever score they are given will eventually become the middle. It's how Elo works.

Says who?

Youre trying to tell me that if the average ELO is say 1300, but if we assign them a starter ELO of say 500, then eventually everyone will sink to that level?

I certainly wouldnt... It actually would probably improve my ELO, as there would be less "derpherding to the PUGzapper" every match, and would make for much better matches.

What it would do is pull all the people that make the game so incredibly frustrating to play out of the mid range ELO and make them earn their spot in the upper ELOs.

[Charons Little Helper]

Kaldor, on 14 April 2014 - 06:18 AM, said:

Says who?

Math does.

For every point in Elo one side gains - the other side loses.

Elo is a zero sum game.

[PhoenixFire55]

Abivard, on 05 April 2014 - 02:36 PM, said:

Notice how most of the white-knights can't even spell Elo correctly, let alone understand it.

PGI spells it "ELO". You expect it to work?

[Kaldor]

Charons Little Helper, on 14 April 2014 - 06:24 AM, said:

Math does.

For every point in Elo one side gains - the other side loses.

Elo is a zero sum game.

Your missing my point. By revising ELO like this, it would separate out the bads and drop them down where they should be. Right now they are being artificially held at a higher ELO because they start so high. The good players will roflstomp the bad players, and hit mid ELO in a short time.

But then again, ELO in this game, is pretty rough, due to small player base, and you being dependent on 11 others which will directly affect your ELO.

This game was much more fun when it was just straight matching....

[Joseph Mallan]

Dock Steward, on 05 April 2014 - 02:46 PM, said:

Funny, after going back through, the only people incorrectly capitalizing Elo (I assume that's what you mean by misspelling?), are the ones who say it doesn't work.

Does it matter? I mean really does the Capitalization really matter? Does ELO really change the whole point of Elo? No, no it doesn't.

[Charons Little Helper]

Kaldor, on 14 April 2014 - 06:31 AM, said:

Your missing my point. By revising ELO like this, it would separate out the bads and drop them down where they should be. Right now they are being artificially held at a higher ELO because they start so high. The good players will roflstomp the bad players, and hit mid ELO in a short time.

Except it wouldn't be mid Elo anymore. It'd be high Elo.

And by your logic - you shouldn't be

Kaldor, on 14 April 2014 - 06:18 AM, said:

"derpherding to the PUGzapper" every match,

but instead you should "roflstomp the bad players" and push your Elo higher.

Your reasoning is inconsistent.

[Kaldor]

Charons Little Helper, on 14 April 2014 - 06:47 AM, said:

Except it wouldn't be mid Elo anymore. It'd be high Elo.

And by your logic - you shouldn't be


but instead you should "roflstomp the bad players" and push your Elo higher.

Your reasoning is inconsistent.

Umm yeah... What your saying makes no sense except your saying that I may be a bad player and you are cherry picking my statement? Complete sentences and thoughts are a good thing...

The fact remains that dropping new players in with experienced mid ELO players is not a good system. It creates a bloated mid section of the ELO populace. New players get farmed because they are getting thrown to the wolves so to speak, but they never really go down in rank far enough to play with players equal their skill and to actually learn the game. It frustrates experienced players who really do want to move up, but are being held back because the majority of the mid ELO populace shouldnt even be there. In 12 man, you W/L is largely NOT determined by you!

The graph for ELO can look how ever you want. Right now I think it looks roughly like an extreme bell curve IIRC, with the mid ELO players and new players creating alot of bloat in the center. It could and should be alot less dramatic than that, with greater numbers at the upper and lower, with far less at the mid point.

[Roadbeer]

Kaldor, on 14 April 2014 - 08:01 AM, said:

The graph for ELO can look how ever you want. Right now I think it looks roughly like an extreme bell curve IIRC, with the mid ELO players and new players creating alot of bloat in the center. It could and should be alot less dramatic than that, with greater numbers at the upper and lower, with far less at the mid point.

Here was PGIs graph on the topic.



IIRC, new players are given an Elo of 1300, which is on the downslope high side of the Elo bellcurve.

Not making commentary at this time, just seemed like this would be helpful to the conversation on the turn it's taken.

[Almond Brown]

Hellen Wheels, on 07 April 2014 - 06:14 PM, said:

There's best solution....

and then there's PGI solution.

Never the twain shall meet.

Would be very interested in how Roland's solution would be enforced? How do you keep out groups of Solo players, that will play well together once the games loads? If you can't enforce it, then it is not a solution at all, just a waste of time trying.

[Charons Little Helper]

Kaldor, on 14 April 2014 - 08:01 AM, said:

Umm yeah... What your saying makes no sense except your saying that I may be a bad player and you are cherry picking my statement? Complete sentences and thoughts are a good thing...

It was a complete sentence - just split by the large quoting system here.

I'm not saying that you're a bad player. I'm saying that by your logic, you're a bad player. (I'm commenting on your faulty logic - not your skill in MWO.)

[Kaldor]

Charons Little Helper, on 14 April 2014 - 08:52 AM, said:

It was a complete sentence - just split by the large quoting system here.

I'm not saying that you're a bad player. I'm saying that by your logic, you're a bad player. (I'm commenting on your faulty logic - not your skill in MWO.)

Then what are your thoughts? Beyond saying my logic is faulty?

But my personal skill, and others can vouch for me, its good enough to easily be in the mid point ELO, heh. Then again, some days Im great, some days it feels like I just suck.

Roadbeer, on 14 April 2014 - 08:13 AM, said:

Here was PGIs graph on the topic.

/snip large annoying jpeg

IIRC, new players are given an Elo of 1300, which is on the downslope high side of the Elo bellcurve.

Not making commentary at this time, just seemed like this would be helpful to the conversation on the turn it's taken.

The upper and lower ends of PGIs graph needs to be fatter to allow more even distribution of the player base.

[Joseph Mallan]

Kaldor, on 14 April 2014 - 09:06 AM, said:

The upper and lower ends of PGIs graph needs to be fatter to allow more even distribution of the player base.

But then it wouldn't be a Bell curve.

[Charons Little Helper]

Roadbeer, on 14 April 2014 - 08:13 AM, said:

Here was PGIs graph on the topic.

Looks like a pretty standard bellcurve.

(Of note - I do agree that MM is far from perfect I do think that the 3/3/3/3 should help considerably - both because tonnage balance will make for more even games - and it'll limit the variables, making Elo steady out faster.)

[Kaldor]

Joseph Mallan, on 14 April 2014 - 09:11 AM, said:

But then it wouldn't be a Bell curve.

Heh, I know what a Bell Curve is Joe. The fact remains that the center portion of it, is far too tall, looking alot like the blue portion of below. We should look more like the red or green line.

Stupid forums: Go here http://upload.wikime...ibution_PDF.svg to see bell curve variations.

The beauty of the Bell Curve is that is doesnt need to zero out until the extreme ends of the scale. For example, the curve Roadbeer posted. The center could be much shorter at about 1300. The extreme ends at 200 and 2200 would still be effectively the upper and lower limits. However the distribution of players from say 500-1000 and 1500-2000 could be much greater, therefore removing some of the bloat from the center of the curve.

Charons Little Helper, on 14 April 2014 - 09:14 AM, said:

Looks like a pretty standard bellcurve.

(Of note - I do agree that MM is far from perfect I do think that the 3/3/3/3 should help considerably - both because tonnage balance will make for more even games - and it'll limit the variables, making Elo steady out faster.)

3/3/3/3 is not really going to fix anything other than maybe be easy to game. Its just another made up limitation PGI wants in place to curb excessive whiny-ness. Time will tell on this one. BV as primary, using a modified ELO system as a secondary, would be better.

[Joseph Mallan]

It really depends on how PGI wanted the numbers to Look Kaldor. The most important part f statistics is what do you want the graph to say!

[Kaldor]

Joseph Mallan, on 14 April 2014 - 09:32 AM, said:

It really depends on how PGI wanted the numbers to Look Kaldor. The most important part f statistics is what do you want the graph to say!

Statistics = lying with numbers!!

[Bilbo]

That particular graph is showing the before and after curve when the fixed a bug shortly after implementation. The darker portion is the after and the intended curve.

[Roadkill]

Joseph Mallan, on 11 April 2014 - 04:42 AM, said:

I guess I don't see this as a problem them RoadKill. Rookies work with vets all the time, in all facets of life, even in recreation. I am someone who is happy there is a better and a worse player than me in every match. Experience and skill are both necessary for a good team.

Context, Joe.

The point is not that there's necessarily a problem. The point is that if there's a problem with matches not being balanced, it's a problem with the matchmaker and not with Elo.

Your Elo rating is just one tool that the matchmaker uses to create matches. If you don't like the matches that are being created and you blame Elo for that, all you're really doing is proving that you don't understand what Elo ratings are.

[Roadkill]

Kaldor, on 14 April 2014 - 06:18 AM, said:

Youre trying to tell me that if the average ELO is say 1300, but if we assign them a starter ELO of say 500, then eventually everyone will sink to that level?

Not exactly. What he's saying is that if you re-define the entry point to be 500, that 500 would then become the average value and all other ratings would re-converge around that point. You'd be re-defining "average" from 1300 to 500.

By definition, the entry point into an Elo system is the average value. It isn't possible to start new players at any other point or it isn't an Elo system.

PGI tries to fiddle with this by adjusting new player Elo ratings by -200 for their first 25 games, but I don't think it has the result that they're hoping it has. It will help a good player's rating jump up quickly, but it will actually make it harder for a bad player's rating to correct itself downward. And all players will have a sudden series of "bad games" after their 25th match as their rating tries to re-adjust to the -200 Elo nerf disappearing.