Forged Alliance Forever Forged Alliance Forever Forums 2014-08-06T20:47:11+02:00 /feed.php?f=2&t=8272 2014-08-06T20:47:11+02:00 2014-08-06T20:47:11+02:00 /viewtopic.php?t=8272&p=78462#p78462 <![CDATA[Re: Rating seasons ?]]>
Poch wrote:
Hi everyone

The best thing to do would probably to have a rating for each map and each spot :D (and you get into the duplicate versions problems and so on)

But i am not trying to raise another controversial discussion about the meaning of rating. Just wanted to talk about the reset thingy. I just wanted to prevent some obvious comment/answers from the first posts^^[/spoiler]


Hey Poch. My take on this is that it would be too complicated to have separate ratings for each map and spot, both to operate and to keep track of. Many player-map-spot combinations would have a really small sample size. If I played Field of the Great Phoenix on the left spot 3 times and lost all of them, what would my rating be? I see a few ways you could alleviate the situation.
Spoiler: show
1.) You could classify all the different spots on each map into different categories, so the sample size wouldn't be as small. If I play field of the great phoenix on the left spot, my rating could be based on "Choke-point big team maps-land spot" or the like.

2.) Expanding on that idea, theoretically, you could alleviate this by having a dynamic rating that, in a given lobby, weights most heavily towards games on that map in that spot, weights second most heavily towards games on that map in a different spot or games in a similar spot category (choke point big team land spot etc.), and finally weighs least heavily towards all other games.

So, on twin rivers air, my rating would be weighted most heavily towards games I've played on twin rivers in the air spot, where I win a ton (say I'm 1700 in that), then towards twin rivers (maybe I'm 1500) and 10x10 team maps where I play air (maybe 1300), then all other games (maybe I'm 1200). So, my rating in that spot would be 1500. But, if I switch to a land spot, it would probably drop to 1300 or 1400. If I join a seton's game, it'll be 1200 in navy, 1000 in air, and 900 in ground etc. because I don't play 20x20 maps as much nor have as much success.

Anyways, like I said above, I think this would all be a bit too complicated and probably hard to do, so I'm not advocating for it, but it's just what I saw as a way to implement the idea.


That said, I do kind of think that if we're making concrete suggestions here, I'd agree with those who'd like the deviation to increase more quickly. Just anecdotally, I notice that ratings don't weight very heavily towards more recent games. I played a ton of 1v1 right when I got the game more than a year ago, then didn't play until about a month ago. My 1v1 rating was way lower than my global rating, and despite winning like 80% of my games since I started playing again it hasn't risen that quickly. I think it'd be better if the ratings would heavily discount games from a really long time ago in favor of more recent games.

Also, I have some friends who feel they are overrated in the game due to whatever circumstances and now have trouble winning, but don't feel their rating going down much. That is much more annecdotal but there you go. So, it'd be nice to see the deviation increase more quickly or some other means to have ratings be more variable. A reset wouldn't be a disaster as far as I see, but I think some people could complain.

Statistics: Posted by sasin — 06 Aug 2014, 20:47


]]>
2014-08-06T20:46:53+02:00 2014-08-06T20:46:53+02:00 /viewtopic.php?t=8272&p=78461#p78461 <![CDATA[Re: Rating seasons ?]]>
Poch wrote:
Hi everyone

The best thing to do would probably to have a rating for each map and each spot :D (and you get into the duplicate versions problems and so on)

But i am not trying to raise another controversial discussion about the meaning of rating. Just wanted to talk about the reset thingy. I just wanted to prevent some obvious comment/answers from the first posts^^[/spoiler]


Hey Poch. My take on this is that it would be too complicated to have separate ratings for each map and spot, both to operate and to keep track of. Many player-map-spot combinations would have a really small sample size. If I played Field of the Great Phoenix on the left spot 3 times and lost all of them, what would my rating be? I see a few ways you could alleviate the situation.
Spoiler: show
1.) You could classify all the different spots on each map into different categories, so the sample size wouldn't be as small. If I play field of the great phoenix on the left spot, my rating could be based on "Choke-point big team maps-land spot" or the like.

2.) Expanding on that idea, theoretically, you could alleviate this by having a dynamic rating that, in a given lobby, weights most heavily towards games on that map in that spot, weights second most heavily towards games on that map in a different spot or games in a similar spot category (choke point big team land spot etc.), and finally weighs least heavily towards all other games.

So, on twin rivers air, my rating would be weighted most heavily towards games I've played on twin rivers in the air spot, where I win a ton (say I'm 1700 in that), then towards twin rivers (maybe I'm 1500) and 10x10 team maps where I play air (maybe 1300), then all other games (maybe I'm 1200). So, my rating in that spot would be 1500. But, if I switch to a land spot, it would probably drop to 1300 or 1400. If I join a seton's game, it'll be 1200 in navy, 1000 in air, and 900 in ground etc. because I don't play 20x20 maps as much nor have as much success.

Anyways, like I said above, I think this would all be a bit too complicated and probably hard to do, so I'm not advocating for it, but it's just what I saw as a way to implement the idea.


That said, I do kind of think that if we're making concrete suggestions here, I'd agree with those who'd like the deviation to increase more quickly. Just anecdotally, I notice that ratings don't weight very heavily towards more recent games. I played a ton of 1v1 right when I got the game more than a year ago, then didn't play until about a month ago. My 1v1 rating was way lower than my global rating, and despite winning like 80% of my games since I started playing again it hasn't risen that quickly. I think it'd be better if the ratings would heavily discount games from a really long time ago in favor of more recent games.

Also, I have some friends who feel they are overrated in the game due to whatever circumstances and now have trouble winning, but don't feel their rating going down much. That is much more annecdotal but there you go. So, it'd be nice to see the deviation increase more quickly or some other means to have ratings be more variable. A reset wouldn't be a disaster as far as I see, but I think some people could complain.

Statistics: Posted by sasin — 06 Aug 2014, 20:46


]]>
2014-08-06T13:30:57+02:00 2014-08-06T13:30:57+02:00 /viewtopic.php?t=8272&p=78440#p78440 <![CDATA[Re: Rating seasons ?]]>
Yeah anyway, even if the community of fa is older than other communities of games, there will be always insults/abuses.
But it's funny I don't remember at all that there were lots of flame/abuses on gpg, and same for balance..

Yes playertrack (it's still up, funny to see the stats, even though I can't get back my replays to see how I was playing), but it was not welcomed at all by everyone in gpg. And even people who didn't want to use it were forced( it just needed one player in the game to have the program on his comp and the game is recorded and you are in it), so nothing you can do against it. But anyway this discussion is pointless, the rating system won't be removed, I'm dreaming.

Sorry Poch for the topic :( I agree for reset or soft reset or deviation increased, for the same reasons.

Statistics: Posted by odin002 — 06 Aug 2014, 13:30


]]>
2014-08-06T10:08:01+02:00 2014-08-06T10:08:01+02:00 /viewtopic.php?t=8272&p=78433#p78433 <![CDATA[Re: Rating seasons ?]]>
In my eyes playertrack introduced out of game flaming with false DNP's and whatnot. In GPGnet some angry chat and ragequits could surely be seen, but that is no less true for FAF.

Statistics: Posted by Col_Walter_Kurtz — 06 Aug 2014, 10:08


]]>
2014-08-06T08:04:40+02:00 2014-08-06T08:04:40+02:00 /viewtopic.php?t=8272&p=78430#p78430 <![CDATA[Re: Rating seasons ?]]>
odin002 wrote:
That's one of the reason why I liked gpg (faf is way better otherwise), no rating system(unless for ladder but whatever, it's way more accurate for this one) in custom games.

And it was working quite well, the good players know each other so they knew if the game was going to be balanced or not, and the beginners were playing together too (because just have to say "beginners" in the game name, or "no good player" in the game title..) .

And thanks to that, people only focused on having FUN, and not flaming a mate because "omg you died at 7 minutes, noob mate..." or some insults.

Yes I disliked playertracker on gpg, because people could flame someone on it (the dnp(do not play) thing). The only good thing about it was the stats on it, this was interesting for someone to see what he is doing well or not.
But at least playertracker didn't say you were 1100, only the % of victory and the number of games, and anyway not everyone was using it.

I would even prefer that the rating system on custom games is removed, like that there will be less flames between people, no crying because one team is 100 point behind, or because it's stacked, and people will only focus on the most important thing of the game : having fun.

Yes the rating can't be 100% accurate and it will be really hard to have a complex system adapted to team games(it won't be done, I agree, too complex to have a rating system which is spot dependant and so on) but tbh I don't feel like the rating system is accurate enough to be used in team games.

The score is only shown in one game(if you are talking about the number of kills/deaths...), not on every game like the rating system.


I think your memory is failing.

Not only most team games on GPGNET were highly unbalanced, but there was a lot more of insults/abuses too.

May I remind you that a guy coded a whole site just to solve that situation (playertrack?).

Team rating is an unbiased evolution of that.

Statistics: Posted by Ze_PilOt — 06 Aug 2014, 08:04


]]>
2014-08-06T00:59:55+02:00 2014-08-06T00:59:55+02:00 /viewtopic.php?t=8272&p=78409#p78409 <![CDATA[Re: Rating seasons ?]]> Statistics: Posted by stalewee — 06 Aug 2014, 00:59


]]>
2014-08-05T20:55:33+02:00 2014-08-05T20:55:33+02:00 /viewtopic.php?t=8272&p=78399#p78399 <![CDATA[Re: Rating seasons ?]]> Statistics: Posted by odin002 — 05 Aug 2014, 20:55


]]>
2014-08-05T20:12:12+02:00 2014-08-05T20:12:12+02:00 /viewtopic.php?t=8272&p=78396#p78396 <![CDATA[Re: Rating seasons ?]]>

And thanks to that, people only focused on having FUN, and not flaming a mate because "omg you died at 7 minutes, noob mate..." or some insults.

Have you ever considered competition might be fun for some people?

Statistics: Posted by Blodir — 05 Aug 2014, 20:12


]]>
2014-08-05T18:10:45+02:00 2014-08-05T18:10:45+02:00 /viewtopic.php?t=8272&p=78392#p78392 <![CDATA[Re: Rating seasons ?]]>
And it was working quite well, the good players know each other so they knew if the game was going to be balanced or not, and the beginners were playing together too (because just have to say "beginners" in the game name, or "no good player" in the game title..) .

And thanks to that, people only focused on having FUN, and not flaming a mate because "omg you died at 7 minutes, noob mate..." or some insults.

Yes I disliked playertracker on gpg, because people could flame someone on it (the dnp(do not play) thing). The only good thing about it was the stats on it, this was interesting for someone to see what he is doing well or not.
But at least playertracker didn't say you were 1100, only the % of victory and the number of games, and anyway not everyone was using it.

I would even prefer that the rating system on custom games is removed, like that there will be less flames between people, no crying because one team is 100 point behind, or because it's stacked, and people will only focus on the most important thing of the game : having fun.

Yes the rating can't be 100% accurate and it will be really hard to have a complex system adapted to team games(it won't be done, I agree, too complex to have a rating system which is spot dependant and so on) but tbh I don't feel like the rating system is accurate enough to be used in team games.

The score is only shown in one game(if you are talking about the number of kills/deaths...), not on every game like the rating system.

Statistics: Posted by odin002 — 05 Aug 2014, 18:10


]]>
2014-08-05T16:24:36+02:00 2014-08-05T16:24:36+02:00 /viewtopic.php?t=8272&p=78386#p78386 <![CDATA[Re: Rating seasons ?]]>
odin002 wrote:
I don't think the rating is adapted to team games anyways sadly, I mean the rating is thought for 1v1.
The rating system for team games should be redone, and at least becoming map dependant and even spot dependant, and ...

Rating should be seen as what it is: a statistical measure of what a player might be capable. So it will never be 100% accurate.
Just think of it. One player of a team has a really bad day while on the other team someone plays one of his better games. That game will be very unbalanced. Can you blame the rating system? No, you cant.

Instead of wasting time to balance games on a very complex rating, where everyone cries if it doesnt fit his expectations and leave, just play more games. This simple rating system is enough to get mostly balanced games.
For example my rating was once >1250 but in team games i announced that i play more like 1050-1100. Normally the host took that into account.

Statistics: Posted by RoLa — 05 Aug 2014, 16:24


]]>
2014-08-05T15:36:22+02:00 2014-08-05T15:36:22+02:00 /viewtopic.php?t=8272&p=78383#p78383 <![CDATA[Re: Rating seasons ?]]>
odin002 wrote:
The rating system for team games should be redone, and at least becoming map dependant and even spot dependant, and also it should depend if you play with people you know(= you already play with them a lot or not) and even depend if you play with a mic or not


I hope you see why this is outrageous and will never work.

A more important question is why do you care if someone got his 2000 rating some way or another? It's the same in every game with score. You can play Metro 64 all day on BF and suck at other maps. It's not like you get free beer for being a level 120. Is it that you leave a lobby when player X with rating Y joins, but instead you should have stayed because he never played map Z and you could have won? If thats the case, you worry too much. In old GPGnet you just played and that had benefits. When they introduced whatsitcalled website with the DNP flag and user rating, things quickly became a spamfest of insults and balance whining. In my modest opinion of course. Not trying to step on any toes just stating what I saw.

Statistics: Posted by Col_Walter_Kurtz — 05 Aug 2014, 15:36


]]>
2014-08-05T13:36:48+02:00 2014-08-05T13:36:48+02:00 /viewtopic.php?t=8272&p=78381#p78381 <![CDATA[Re: Rating seasons ?]]>
The rating system for team games should be redone, and at least becoming map dependant and even spot dependant, and also it should depend if you play with people you know(= you already play with them a lot or not) and even depend if you play with a mic or not, in fact there are so many factors to take into account lol...
t's not easy to make a good rating system.. =/

For example when someone only plays one map and has a high rating, he doesn't deserve this rating on other maps where he is bad or at least not as good .
What about the spot too, on some maps there is a spot for air, and spots for land, what about people who only play the land spot or air spot ... Same issue.
and so on..and so on...

And why not a 1v1 reset rating too?

Statistics: Posted by odin002 — 05 Aug 2014, 13:36


]]>
2014-08-05T11:03:22+02:00 2014-08-05T11:03:22+02:00 /viewtopic.php?t=8272&p=78370#p78370 <![CDATA[Re: Rating seasons ?]]>
I'm new to the FAF rating system. I have 15 rated games at the moment.

Anyway resetting stats generally creates a lot ot outrage. In fact it's used as a punishment for cheating (Battlefield for example).

Statistics: Posted by Col_Walter_Kurtz — 05 Aug 2014, 11:03


]]>
2014-07-31T17:11:56+02:00 2014-07-31T17:11:56+02:00 /viewtopic.php?t=8272&p=77948#p77948 <![CDATA[Re: Rating seasons ?]]> Statistics: Posted by Meteora — 31 Jul 2014, 17:11


]]>
2014-07-31T16:34:17+02:00 2014-07-31T16:34:17+02:00 /viewtopic.php?t=8272&p=77945#p77945 <![CDATA[Re: Rating seasons ?]]>
I think it'd only help more players, play with people they don't normally - and increase the average rating of the entire community. I suggest a rating reset every 6 months. Maybe keep 1v1 the same for those die-hards :P

Statistics: Posted by VoiceofReason — 31 Jul 2014, 16:34


]]>