if you look at realtime ladder rating, you can see it worked better in this 6 months.
The noexpire thing was the key for prevent stall. Also it incentive to play more games.

The noexpire thing was the key for prevent stall

It's hard to stall with 5 min autoboot...

you misunderstand me.
I meant to say, that you could try to make a good run of win in row and w8 for the other loss expire. Then start to play again.
The no expire and the fact you have to play at least one game every 3 days prevent someone jump 1st without deserve it.

But multiday ladder can't work like RT ladder - you can't force people to finish one game every 3 days with 3 day boot. And if you just change rating system to TS, it won't solve stalling problem. We don't know(at least i don't know)how fast games are 'faded out'.

Imagine this:
Player A starts with 5 games at once. He wins 10 games but ultimately he stalls 4 of 5 to prevent breaking winning streak. He wins 15 in a row, then he surrender all 5 games.
Then he tries another run - he wins 25 in a row (with stalling), so his history looks this way:
15 wins followed by 5 loses followed by 25 wins.
The question is - how big is influence of those 5 loses.
If it's marginal - you can still stall to get good ladder result.
If it's big enough and after 25 games your old loses are still affecting your rating - read piggy's post. You may imagine a lot of players with alt accounts trying to get as good result as they can with fresh start, then slowing game pace to the lowest possible.

As a anti-stalling mechanism i would propose one more my idea:
allow to commit only your longest awaiting ladder games - you can't stall with such restriction
(maybe with exception for one/two games to reduce inconvenience - so you can commit game only if no more than one/two others game has longer awaiting time).

@kryzhu.
Look at realtime ladder now.
You have to finish at least 50 games for be competetive, otherwise the variaton is go high you will never go to top of ladder.
After 50 games, one more loss will make you lose 10/15 points. No point for stalling. Also stall in such system is useless. In the opposite, more games you finish, more harder will be increase your rating (or decrease ofc).
If you win 15 games you will have 145 as variation more or less.
let's say you do very good and you have 2500 trueskill meant (that is A LOT).
2500 - 3*145 =2070. Never enough for take top.
So problem solved.

So, if i won 15 next games against you you would still be higer rated than me. I don't think this is what we expect from ladder ranking, as ut should represent contemporary results.

Why has nobody addressed Hennns suggestion? I think it's an excellent idea. A static rating system based on initial rating of players going in. It's fair in Chess and it can be fair here. A suggestion based on this system would be to have a looser matching system - with constant ELO rating, after a couple of months, we have a true ladder. Something like the RT ladder.

@Krzychu. NO. We expect a FAIR ladder. Gnuff's idea makes sense. His system would give the Trueskill result of all players, ranking them accordingly. I still prefer a dual system or true ELO though.

Edited 8/29/2014 11:33:33

Till you get variation < 80, you have no chance to be in top of ladder.
This prevent run of 15 games. At same time prevent stall as well. Bc you if you surr you will not lose much points.
Look at his balls in realtime ladder. HE was 1400, now he is 1900. Past lost count A LOT less than new wins.
Considering no expire, after some months (2,3?) we will all have variation low (under 80) , nobody gain nothnig to stall, nobody can make run and jump 1st. Only skill will help you to do well.
No way to cheat, also bc system pick opponent for you (ofc in a range of 20/30% of your true skill meant).

It will be much more interesting. When a new Number 1 emerges, it will be a notable event and not the result of some run.

yeah for this i want that the same rating system from realtime ladder will be copy for each ladders. Adding for the matchup 30% (or 20%) of your trueskill (not rating!!!).

30% I think. The wider the better imo (to a point).

Look at his balls in realtime ladder. HE was 1400, now he is 1900. Past lost count A LOT less than new wins.
Considering no expire, after some months (2,3?) we will all have variation low (under 80) , nobody gain nothnig to stall, nobody can make run and jump 1st. Only skill will help you to do well.

What will be his rating if in he wins next 45 out of 50 games?
And what it would be if he wins next 45 out of 50 games with new account? We can only specalute without exact formulas or simulation, but is surely be higer in second case.
So 'no-expire' is invitation to make ladder a 'alt-arena'.

Considering that you will have to play with player with yuor similar skill, i say is not possible win more than 80% in long run.
However you may be right. But is Always better this system than the 1v1 ladder system.

Alt is Always a problem anyway. And the system we use atm is worse about alt. :)

@Krzychu. NO. We expect a FAIR ladder

You have played 300 ladder games during last two years. Imagine now you become the best player in the world. You win all your games in the next 6 months(75, assuming similar pace). But well - you are ranked 10th, because there are few slyboots, they've created new accounts and won 35out of 40. That's fair?

Edited 8/29/2014 12:09:15

no way you win 30 out 40 and you are 1st.
You need to play at least 70 games for have a varation low enough for be 1st.
Consider that in 1v1 ladder more player will join than rt ladder.
however, you rather keep the system we have atm in 1v1 ladder???

Edited 8/29/2014 12:06:33

@Gnuff

And the system we use atm is worse about alt. :)

[Argumentum ad personam mode on]That's why you have 3 accounts in RT ladder, but 2 with worse record are abandoned:P[Argumentum ad personam mode off]
More seriously: yes, it is, nobody claim it's good, but simple switch from bELO to TS isn't good solution


And can you explain me how to calculate variation for X games ?

Edited 8/29/2014 12:10:57

As I said before, not Hennns, I think that chess ELO could be a solution. The problems of a dynamic rating would dissapear too.
With chess elo if you want to go up in the ladder you need to win against higher oponents. Even if you are matched with players with less level, the rating variations are ridiculous when you win, so you NEED to win against the good ones, no lucky strikes to avoid top players can make you nº1. Stalling would be meaningless too because the rating variation is determined by the elo at the start of the game so you will lose the exact same amount of rating losing in 1 day or 2 month later, not taking into account your rating variation in other games.
One game can only modify your rating, at most, the same as K value (usually 15), so a lucky win against a top player isn't enough. Same for loses, if you have a bad day against a bad player that doesn't mean you will drop lots of position from the ladder. Only one game has little impact.

Sorry Memele I don't know why I called you Hennns :/

I agree totally.

is already a while since the creation of this thread. When Fizzer you will make the polls? :)

^^^I was just about to say that