I would like to hear more about Memele/Hennes comment on the K factor. Is the chess variation of ELO better than TS?

My main concern right now with TS on the RT ladder is it seems to heavily favor volume of games.

Edited 9/2/2014 17:25:08

Hennns*

The Elo used in chess is designed for 1v1 games. The rating you have is used to predict the outcome against your opponent. The K-factor decides how much a rating can change based on a single game. ex) if K=15, you could go max up 15 points (or down), regardless of your opponents rating. the K value can vary based on several things. for example higher rated players could have lower k values, making their rating more stable. (Something like this would be very nice for wl imo).

Trueskill is a modified version of the Elo rating system made by microsoft, it is designed to be used for Xbox. It does not have a K-factor and works slightly different. And that's one of the reasons I think the Elo rating system is better suited for warlight.


You can read more on wiki:
http://en.wikipedia.org/wiki/Elo_rating_system
http://en.wikipedia.org/wiki/TrueSkill

@ChrisCMU
Almost all rating systems favors volume of games, that's something very difficult to correct. Obviously there is systems that favors it more than others.

In chess elo, your variation is at most K each game (usually the higher the ranking the lower the K). If you play a lot and keep winning, your rating will increase but then some of these options will happen:
1) Your rating grow but your opponents remains the same (i.e, if you are in the top), so each victory gives you less points (and each lose its a bigger elo loss). IF you keep winning for a long time (almost impossible) your wins will be 1 point and losses -14 (K=15). IF someone manages to keep winning ELO despite of this, that means he deserves it.

2) Your rating grow and you start playing better oponents. If you keep winning your rating growing was correct. If you start to lose you will return to your level so nothing wrong.

vote guys! VOTE!!!!!

i wish the Poland template submitted to the realtime ladder wasn't using sze's "strategic" settings, but the normal ones. Poland is a really interesting map, but those settings give no room for error on picks or expansion / countering strategy, you need to be very experienced in the map and in strategic low luck settings in general to win it, slaughterhouse for players below the top 20.

I gave it a generous 1.

I've always wondered, why 351% for the blockade card (in 1v1)?

VOTE!!

I've always wondered, why 351% for the blockade card (in 1v1)?

if you mean why 351% and not 350% because you think both are the same, then you are wrong.
if you make a blockade of 3 and it's 350%, the blockade will be 10 (NOT 11, yes I know that 10.5 should be rounded up but it's not like that. it is rounded down). however, if it's 351% then the blockade will be 11.

if you meant why 351% and not any other number (let's say 300% or 500%), then it's just what fizzer decided when he made the template. it's like asking why 3 starting positions and not 4 in the strategic 1v1.

Edited 9/5/2014 09:58:19

as for your example about starting spots, i think it has a reason:

With even number of starts, person with first picks could get full intel aswell, while person with 2nd pick might be doomed twice (miss most important pick and opponent knows everything)

When do the polls close?

about weighting games:

Also, games never expire like they do in the other ladders, but the TrueSkill algorithm weights newer games more highly, so you still have the ability to move your rating over time.


from http://blog.warlight.net/index.php/2014/03/website-update-2-5-real-time-ladder/


about alts:
games never expire -> you can't get rid of your old loses -> it's easier to play with new account without loses than recover from lower rating
Gnuff was angry that i mentioned this in another thread and probably will be again, but i can't do anything about that, as he is excellent example why i don't like TS. He already played RT ladder with (at least) 3 accounts. Two of them are already abandoned - no new game for a while(Gnuff and Killua). He play only with Marquis now. And new question of the day - compare a bunch of first games of all those accounts and tell me which has most wins.
There will be a lot of such behaviour if we use TS for 1v1 ladder.

Anyway - i found that:
http://blog.warlight.net/index.php/2012/01/trueskill/
there's a link for WLTrueSkill, but this file is not avaialble now. Maybe fizzer could reupload it, as it would be nice to do some simulations. (including end date for games API would be helpful too:P)

Edit: starting over is only temporary solution, there are still new people joining and players improve over time.

Completely agree with this. Games need to at some point expire or alts become an even bigger problem.

i wish the Poland template submitted to the realtime ladder wasn't using sze's "strategic" settings, but the normal ones. Poland is a really interesting map, but those settings give no room for error on picks or expansion / countering strategy, you need to be very experienced in the map and in strategic low luck settings in general to win it, slaughterhouse for players below the top 20.

Would you care to submit another Poland 1v1 template? My initial intent was to eliminate luck as much as possible, however I still think many decision in rt ladder would be based on instinct. On Poland template you need to calculate far too much for 5 minutes to be enough.

Completely agree with this. Games need to at some point expire or alts become an even bigger problem.

I agree that what gnuff did is not fair, but I don't agree with the point you and Krzychu are making. How exactly expiration prevents alt abuse or how does it discourage?

Just to be clear, as my name is always done, i want say that most of opponent use more than one acc as well. Hhh timi TWM gui potato and many others.
Expiration make 0 sense though.

I haven't been keeping attention but surely the most simple idea would be best. Either:
•Increase amount of games needed to rank - to 20
•Decrease the amount of games you can play at a time to 3
This would pretty much limit a players chance of stalling. Though only 3 games at a time could effect players that don't stall.

if you limit to 3 games at a time, please only during the first 20 games :)

3 games at time are pretty low though. If you get some opponent like math wolf you will never get a rank xD

I would like to increase the max games to 10, a limit of 3 would be a real punishment for me. I would join the ladder with more alts, just to play more games

I agree with MOD. We all seem to focus on the alt/stalling of the ladder. While both are 'problems', there are over 250 1v1 players and around 200 2v2 players. A very small but very visible and vocal portion are stallers and alts. I have an alt on the ladder just so I can get additional games. Both accounts are ranked the exact same and have close to 100 unexpired games. There is no need to punish the many for the acts of a few. Fizzer is being asked to fix a problem that only exists for 10% of the ladder, which might even be generous. The discussion should revolve around how to get more players in the ladder playing at a higher level and keeping the ladder from monotonous clicking and calculating.

Qi had a solution to this a while ago, which was cap the number of ladder games you can play at different amounts depending on how many unexpired games you have.

Something like 2 for 0-5 completed games, 3 for 5-10, 4 for 10-15, 5 for 15-20, and unlimited for 20+. I don't know what the optimal limit would be for each level of number of games, but I think something like this would help solve both of those problems.