I know we've been talking about this a lot since the 3 month expiration timer started kicking in but based on everything that I've seen happen over the last few weeks, I'd like to throw my lot in with the fact that I no longer believe that games should be permanently taken off of your record after 3 months (unless we move to a [**Seasonal Ladder option**](http://warlight.uservoice.com/forums/77051-warlight-features/suggestions/1521999-ladder-seasons-?ref=title) where we have a continually running Ladder and then a 3 month total reset stat tracking calculated separately). I think the current format is still serviceable but I'd like to then see expired games raised to maybe 10%-25% of the original value and not eliminated entirely. What are everyone else's most current thoughts?

I agree - As it stands, your ladder ranking is both a reflection of your skill AND a measure of your activity, but since the matches are set up based on the resultant scores, it seems like someone not playing very many games or taking a hiatus shouldn't really come into play there, since generally someone is going to be just as skilled now as they were before they took a two week break. It also makes the scores more volatile, susceptible to large variations due to winning or losing streaks and the timing of their occurence. I think it would be nice though to have the ladder go for a specific time and then have everyone's slate wiped clean. There could be a list of previous winners of each ladder "season" so as to give them due credit :)

Wouldn't that cause even more stalling by the top players nearing the end of the period to start a new "season"? Also what would happen to the games that were in play at the time?

I am actually coming around to the idea of altering the expiration window. I've reduced my number of games lately and can see that the expiration of big sets of old games are going to have an effect on my rating, mostly positive.

However, I am not sold on the "reduced weight" mechanism. Are there other ways to tweak the system? For example, instead of the games expiring based on the calendar, they would expire when a certain number of games have been played. With this system, the last X number of games (say 100) are used for your rating, regardless of how old they are. Only the most recent 100 are used and if you play more you can bump ones off the back end.

I suspect that this runs into problems with the rating system because a game might be expired for one person and not the other person in the game, so possibly this has a fatal flaw to it.

Eitz, what about a phased reduced weight?
0-3 months 100%
3-6 months 50%
6-9 months 25%
9+ months 0% (or some other low level)

Spreading out the length of the expiration will smooth out the shifts in ratings quite a bit.

NoZone: I like your idea of using the most recent 100 games as a ranking **A LOT**! As a secondary solution, I like your % post as well but this most recent 100 games idea is the best solution I've heard yet.

Devilnis: That's pretty much what the Seasonal Ladder idea is all about. [Check it out here](http://warlight.uservoice.com/forums/77051-warlight-features/suggestions/1521999-ladder-seasons-?ref=title).

I would lend my support to any of these ideas. Although the seasonal option is the most appealing to me because it would allow for a lifetime acknowledgement of wins, I also think it would be the most succeptiable to various problems that would divdie warlighters (I.E how many games you need to participate in to be eligiable? Or when a new player should be allowed to enter the ladder, during or only at the beginging of a season? Or even simply a lack of interest once you have lost a few games in a season and would not be able to reach the #1 spot).

Knowing this, I think the reduced weighting system would be best, albeit with much smaller drops in percentages from period to period.

Something else I just noticed.

Amongst several other concerns that have already been stated by others, I wanna give an example of why the current 3month system is far too short a time for games to expire in.
Despite playing at a high clip since the start of the 2v2 ladder (currently set to 5 games at a time) Ragingpikey and I have only managed to complete 29 games, 11 of which are now expired. This just goes to show that even with two active players playing at a good pace, the speed of the 2v2 ladder is one which can barely outrun the expiration of its games, as it is we only sit 8 games up on the 10 game minimum.

Oh and incase I haven't griped enough already, we have a far superior overall winning percentage then the current usurpers of our throne Eitz/Bp...Just saying :p

I like the phased idea, but I propose a better solution, a game's weight in the calculation of your overall score would be :

if(days since completion<=90)
weight=1
else
days=days since completion - 90
weight=(1/(1+(days/10))

sorry for the psudo-code but It seemed like the best way. If you don't understand this, it means that up to 90 days a game counts 100%, at 100 days it counts as 50%, after 110 days it counts as 33%, after 120 days it would count 25% and so on

+1 to any of the above.

Honestly....i happen to like what has happened to the ladder with the 3 month shelf life. Look at it now, it has new life compared to what it was. People are now trying desperately to get their rank back up because everything is expiring. It's pretty exciting for the ranks of 15-50 because now they're getting a shot at being in the top 10. I'm finding that the ladder is more exciting then ever.

Now, i do agree with Eitz that a Season would be nice, but this on going one where you have to keep at it is really fun.

R - That will happen almost as much with weighting. The difference between counting for nothing and counting for 10% is not a lot in terms of overall rank.

Out of 14 teams currently ranked on the ladder, there is not one with more then 21 games played, instead most are hovering just above the minimum of 10.

Despite Pikey and I playing in 5 games at a time for close to two months now, we still only have 14 ranked games counting towards our stats. Of course it would be closer to 15 games if we hadn't been waiting for the last 30 days for Dazed and Insane to finish his vacation time (Thats another issue entirely though).

Although I may not agree with the 3 month expiration time on the 1v1 ladder, at least there a negative effect on the sample size of the stats can be overcome by simply playing quickly. To do this in the 2v2 ladder would be unrealistic, especially when taking into account possible delays from vacation mode.

Sorry if my post was a bit repetitive of my earlier ones, but I feel like this issue needs to be addressed more than it has.

I agree it makes sense for the 2v2 ladder to have a longer shelf life than the 1v1. I'll add this to my list of things for the next release.

Thanks for the response FIzzer

I am finding that if i were to just not play or leave the ladder and wait for some of my large blocks of losses to go away i would go up in rank a fair bit. In fact it seems the majority of the player ranked above me do not have more than one page of active games,I think this 3 month expiration is encouraging people to not play as much. i lowered my game count to one for the summer but there isnt much to encourage me to raise it when my ranking will just go up the less i play. we all have to start somewhere and if you were to go back to my first games on the ladder it would show i have just as much if not more to lose by having those games count i just feel that it is a more accurate representation of skill not just a way to manipulate the system

|> *I am finding that if i were to just not play or leave the ladder and wait for some of my large blocks of losses to go away i would go up in rank a fair bit*

That means you're playing better recently than you did in the past - congrats!

|> *In fact it seems the majority of the player ranked above me do not have more than one page of active games,I think this 3 month expiration is encouraging people to not play as much.*

[Correlation does not imply causation](http://en.wikipedia.org/wiki/Correlation_does_not_imply_causation). This is a common mistake, but you need to realize there are countless other factors in play that could explain the correlation.

In fact, the ladder formulas can prove that playing more games doesn't hurt your rating - your overall wins and losses and whom they're against is the only thing that affects it. Having fewer games makes your rank more volatile, and you're only examining those at the top so you're only noticing the upsides of that volatility. You would need to do a statistical analysis of every player to make such a statement, and if you do I believe that you'd find it's not true.

Your right im sure, but if for me it helps to not play as much im sure its doing the same for others. My concern is that if i were to get lucky and beat the impaller and leave the ladder right then and wait until the 3 months comes up to my recent win streak i would be a top ten player perhaps, and i do not feel that would be accurate. I know this wouldnt effect anything in the long run but i do think it would lead to that person playing as little as possible to keep there score as high as possible.

I understand that in the end this will all even out but i do think it is perhaps slowing the ladder down a bit and is a way to manipulate your score.
I read all the posts and think there have been some good ideas already said that would fix this

A weighted shelf life would be interesting. I doubt the shelf life affects the playing habits of players rated 1900+ though: there is a noticeable gap in play of players rated 1900+ and everyone else, so they don't need to just stop playing games to remain the best.

Two things I think would be interesting:

1. Ladder challenges: Players can invite/challenge other players to ladder matches. And these challenges could be accepted or declined. For example, if I am in 10th place and want to move up, I could just challenge the guy(s) in front of me. This is how championship boxers operate.

2. Instead of a seasonal ladder, a seasonal ladder tournament: A seasonal ladder doesn't make sense (skill ratings are supposed to measure the present, not individual seasons). But to have a seasonal tournament of the top 32 or 64 players would be cool. Such games could replace the random ladder matches. A lot of good players seem to leave the ladder(s) once they achieve a personal goal (or realize they can't dethrone the Impaller? lol). A tournament might keep such players involved (new goal, new challenge).

ladder challanges could be used to inflate your score...