Elim - Changes & New Season Coming!

[Wirah]

I wonder if you can buy different ship skins in elim using pub-bux. 3 dif skins for jav/ 3 for warbird. I had this idea for TWD but maybe since they added spacebux in elim, they could try it out?

Buy ship 3 (Blue jav): Cost 100k spacebux + 40 jav elim games played
Buy ship 4 (cool design jav): Cost 250k spacebux + 80 jav elim games played.
Buy ship 5 (USA jav): Cost 400k spacebux + 120 jav elim games played


Same for the warbird- but with ships 6-8.

This idea would mean spacebux will be worth collecting/ people with the usa skin ships might impress other players. Could make elim more popular.

Like this idea. Flying a golden warbird/jav because you're #1 sounds pretty awesome too. Silver for top 5.

[Rab]

Yeah that's fun. Hunt the shiny jav !!

[Dierienow]

I think it's a good idea too. I'm sure there is a reason why it can't be done though, even if that reason is as simple as the fact that there is only 1 developer and they've given everything to this project already. Would be great to see though, and is one of my SS2 dream features. The ship skins that is.

[ForgottenGhost]

found a bit of a bug:

if someone specs (or lags out) 3 seconds into a game, the bot seems to treat that like a legit elimination, and therefore nobody else can join because someone changed their mind about playing before a single kill/death occurred

[ForgottenGhost]

Update: the bug I mentioned above (regarding lagout eliminations) still has not been rectified 2 weeks later.

Also, my two cents on the ratings:

1). Once you hit 100% confidence it seems nearly impossible to move through the rankings, it slows to a complete crawl, and by then you aren't truly rewarded for being active, or even winning. For example you could win 3 games in a row against 7 people, and you might only move up 1 rating per round despite being ranked #40 or something. When the people in front of you have 700 more rating, it feels quite pointless.

2). I've seen some of the higher ranked players like Steadman WIN elim games, and then LOSE rating points, which seems a bit broken.

[qan]

Re: bug, yeah, mild RSI/carpal tunnel acting up (do your stretches, kids), haven't been coding or typing as much to give the wrist a rest. I checked it out before and it didn't immediately pop out. But, found the cause I believe, should be fixed after next reboot.

1) That's known as ELO hell (roughly speaking ... it's slightly different in our case). There's not much that can be done about it, except maybe having 100% confidence reached at a higher number of games, or having 100% confidence calculated based on the median of the most active 20 or so players. (Only difficulty there is that confidence is calculated on a logarithmic rather than linear scale currently. Could go back to linear, though it's quite janky for low # of games played.) That would possibly reward activity over skill. It's hard to find the perfect balance. We are also looking at ratings decay for inactive players for next season. And I still need to add the one week cutoff, where you just aren't on the ladder if you're inactive for one week.

2) Unfortunately even going 10-0 1v1 against a low-rated player will lose you rating if you are a well-rated player. This is inherent in almost any formula that folds the skill of the enemy into it. We could look at basing the rating less on the skill of the players involved. That does make it more open to being gamed, though. Other games deal with this by matchmaking, so that Steadman just wouldn't get matched against players who are so far beneath his skill. But we don't have that luxury. One plan to offset this is to make streaks exponential in terms of reward. Haven't played enough with the numbers though. Bottom line is we'd really need to get away from average enemy rating to fix this, which opens up the problem of quite easily gaming/feeding ratings during the slow times.

There's a link to the editable ratings spreadsheet in !rh if you want to play with numbers a bit.

[ForgottenGhost]

That's known as ELO hell (roughly speaking ... it's slightly different in our case). There's not much that can be done about it, except maybe having 100% confidence reached at a higher number of games, or having 100% confidence calculated based on the median of the most active 20 or so players. (Only difficulty there is that confidence is calculated on a logarithmic rather than linear scale currently. Could go back to linear, though it's quite janky for low # of games played.) That would possibly reward activity over skill. It's hard to find the perfect balance. We are also looking at ratings decay for inactive players for next season. And I still need to add the one week cutoff, where you just aren't on the ladder if you're inactive for one week.

I'm not against the linear scale idea, but I wasn't necessarily suggesting it need be COMPLETELY activity based either. If you're a great player, I don't see why you should have to prove you are over and over again constantly. Also, my issue isn't necessarily that it's too skill based, it's that it's specifically too skill based for the games that you build to the 100% confidence rating, and then not truly counting the games played afterwards.

In terms of the Elim "season" there's already almost no reason to play if you're already at your 100% confidence the way it is now. For example, if Zidane never plays another Elim match, he will likely win this season completely unchallenged for Javelin, except maybe by the person ranked #2.
On the flip side a player like me who is ranked lower, has a 0% chance of even breaking into the top 10, even if i'm extremely active, and the "winningest" player of all time.
It's not necessarily a bad thing, but those needles aren't moving.


If that's the way it was meant to be designed, feel free to ignore this message. But then the Elim season might as well only be a couple weeks long rather than 2-3 months.
Otherwise, I think there needs to be slightly more incentive to keep playing for the entire season, and I already see the numbers heavily dwindling.