Upvote Upvoted 50 Downvote Downvoted
1 ⋅⋅ 5 6 7 8 9 10
RGL PUGs Public Alpha
posted in Projects
211
#211
6 Frags +
LastandFirstSo, I think when there are more than 2 pugs running, the balancing is worst. Like a mix of fat kids needing picked and less players to choose from in the pool, and you've got a dog shit pug. Maybe some sort of high and low open pugs to try and keep the skill level somewhat fair. If one pug needs additional players people can be brought up or down. Idk how hard that would be though.

No. There's no reason for a pug to be so unfavorable when the auto-picker just has to shuffle some people around.

https://i.imgur.com/ijWjSst.png

12min pug btw

[quote=LastandFirst]So, I think when there are more than 2 pugs running, the balancing is worst. Like a mix of fat kids needing picked and less players to choose from in the pool, and you've got a dog shit pug. Maybe some sort of high and low open pugs to try and keep the skill level somewhat fair. If one pug needs additional players people can be brought up or down. Idk how hard that would be though.[/quote]
No. There's no reason for a pug to be so unfavorable when the auto-picker just has to shuffle some people around. [img]https://i.imgur.com/ijWjSst.png[/img]
12min pug btw
212
#212
15 Frags +

the new eu pug service has a system where you can vote for which captains you'd like before the pug starts if theres more than two (optional vote) i think thats cool imo

the new eu pug service has a system where you can vote for which captains you'd like before the pug starts if theres more than two (optional vote) i think thats cool imo
213
#213
9 Frags +

re: last 2 weeks of elo discussion: people underestimate how large of a playerpool you need for elo to actually make sense. 27 people added up for 12 player games is just not enough - elo is an illusion at that scale. in my naive opinion. An RL algorithm might be able to, over time, begin picking teams that edge closer to ties (you'd need a bajillion matches to train it though), but a simple elo/glicko algorithm is ineffective at this scale.

bug report: last night fyg was added up to both open and captain pugs, started picking a captain pug, but then the open pug started and fyg got picked by the bot when they shouldve been removed from the queue after captain pug starting picking.

re: last 2 weeks of elo discussion: people underestimate how large of a playerpool you need for elo to actually make sense. 27 people added up for 12 player games is just not enough - elo is an illusion at that scale. in my naive opinion. An RL algorithm might be able to, over time, begin picking teams that edge closer to ties (you'd need a bajillion matches to train it though), but a simple elo/glicko algorithm is ineffective at this scale.

bug report: last night fyg was added up to both open and captain pugs, started picking a captain pug, but then the open pug started and fyg got picked by the bot when they shouldve been removed from the queue after captain pug starting picking.
214
#214
4 Frags +
MozzarellaFireFoxre: last 2 weeks of elo discussion: people underestimate how large of a playerpool you need for elo to actually make sense. 27 people added up for 12 player games is just not enough - elo is an illusion at that scale. in my naive opinion. An RL algorithm might be able to, over time, begin picking teams that edge closer to ties (you'd need a bajillion matches to train it though), but a simple elo/glicko algorithm is ineffective at this scale.

The Elo/glicko system is a rating system, which is different from the matchmaking system (that may use Elo). I'm also pretty sure the number of played games per player is a better indicator than how many people add up. There are other challenges with small number of people adding up but not having to do with Elo.

[quote=MozzarellaFireFox]re: last 2 weeks of elo discussion: people underestimate how large of a playerpool you need for elo to actually make sense. 27 people added up for 12 player games is just not enough - elo is an illusion at that scale. in my naive opinion. An RL algorithm might be able to, over time, begin picking teams that edge closer to ties (you'd need a bajillion matches to train it though), but a simple elo/glicko algorithm is ineffective at this scale.[/quote]

The Elo/glicko system is a rating system, which is different from the matchmaking system (that may use Elo). I'm also pretty sure the number of played games per player is a better indicator than how many people add up. There are other challenges with small number of people adding up but not having to do with Elo.
215
#215
4 Frags +

can there be a cooldown on bans, ive subbed myself maybe 4 times over 6 months and im sitting on a day and a half ban. it would be cool if at least subbing yourself out would have a reset on ban duration every month or so. I get its to discourage people from pug jumping but its kind of absurd

can there be a cooldown on bans, ive subbed myself maybe 4 times over 6 months and im sitting on a day and a half ban. it would be cool if at least subbing yourself out would have a reset on ban duration every month or so. I get its to discourage people from pug jumping but its kind of absurd
216
#216
2 Frags +

Speaking of which, subbing / reporting a player also results in an error page (don't recall specifics offhand; it's fixed by refreshing)

Speaking of which, subbing / reporting a player also results in an error page (don't recall specifics offhand; it's fixed by refreshing)
217
#217
-5 Frags +

Can we stop allowing players that have never played a single RGL season or any season in any comp league from playing pugs? Its extremely frustrating. I get its open pugs but come on bro you are asking for people to sign up on alt and just use cheats. Before you say just add up captain pugs, it is nigh impossible to get those going because for someone reason everyone think its reserved for the highest level players.

Can we stop allowing players that have never played a single RGL season or any season in any comp league from playing pugs? Its extremely frustrating. I get its open pugs but come on bro you are asking for people to sign up on alt and just use cheats. Before you say just add up captain pugs, it is nigh impossible to get those going because for someone reason everyone think its reserved for the highest level players.
218
#218
-4 Frags +
Prime-SanityCan we stop allowing players that have never played a single RGL season or any season in any comp league from playing pugs? Its extremely frustrating. I get its open pugs but come on bro you are asking for people to sign up on alt and just use cheats. Before you say just add up captain pugs, it is nigh impossible to get those going because for someone reason everyone think its reserved for the highest level players.

literally all I can think of as a solution is to have a newbie category and have to play a certain amount of games to go through until they are allowed in open/captain pugs, let's say random number 300-500 games or at least a season, to get the gist if not fundamentals of 6s/highlander.

any other solutions or thoughts?

[quote=Prime-Sanity]Can we stop allowing players that have never played a single RGL season or any season in any comp league from playing pugs? Its extremely frustrating. I get its open pugs but come on bro you are asking for people to sign up on alt and just use cheats. Before you say just add up captain pugs, it is nigh impossible to get those going because for someone reason everyone think its reserved for the highest level players.[/quote]
literally all I can think of as a solution is to have a newbie category and have to play a certain amount of games to go through until they are allowed in open/captain pugs, let's say random number 300-500 games or at least a season, to get the gist if not fundamentals of 6s/highlander.

any other solutions or thoughts?
219
#219
22 Frags +
Prime-SanityCan we stop allowing players that have never played a single RGL season or any season in any comp league from playing pugs? Its extremely frustrating.

Open pugs are just fine. Asking new players to sign up for a full season of 6s or hl before they can play pugs just adds massive barrier of entry. NA Tf2's pug system is probably in it's best shape since pug champ era, and having an open pug system where new players can simply add up and play 6s/hl for the first time without having to commit to a full season is a good step for RGL. (not to mention it avoids the biannual "wheres this months pug discord at?" tftv/reddit thread).

Prime-SanityI get its open pugs but come on bro you are asking for people to sign up on alt and just use cheats.

If this is your primary concern, rather than being annoyed at noobs on your team, then this post makes more sense. If someone decides to make an alt and cheat, that sucks for but its a worthwhile trade off if it means more new players have access to pugs.

Prime-SanityBefore you say just add up captain pugs, it is nigh impossible to get those going because for someone reason everyone think its reserved for the highest level players.

Be The Change.

[quote=Prime-Sanity]Can we stop allowing players that have never played a single RGL season or any season in any comp league from playing pugs? Its extremely frustrating.[/quote]
Open pugs are just fine. Asking new players to sign up for a full season of 6s or hl before they can play pugs just adds massive barrier of entry. NA Tf2's pug system is probably in it's best shape since pug champ era, and having an open pug system where new players can simply add up and play 6s/hl for the first time without having to commit to a full season is a good step for RGL. (not to mention it avoids the biannual "wheres this months pug discord at?" tftv/reddit thread).
[quote=Prime-Sanity]I get its open pugs but come on bro you are asking for people to sign up on alt and just use cheats.[/quote]
If this is your primary concern, rather than being annoyed at noobs on your team, then this post makes more sense. If someone decides to make an alt and cheat, that sucks for but its a worthwhile trade off if it means more new players have access to pugs.
[quote=Prime-Sanity]Before you say just add up captain pugs, it is nigh impossible to get those going because for someone reason everyone think its reserved for the highest level players.[/quote]
Be The Change.
220
#220
1 Frags +

i think open pugs are completely fine and i can only think of one blatant alt that ive seen playing RGL pugs and i think he is banned now anyway. i think the only reason open pugs can be bad sometimes is when theres a big disparity in skill between the players added up and creates uneven pugs, but thats better than never getting to play at all honestly. im also very glad most of the adv/inv players are playing captain pugs now, because getting raged at by higher level players was so unfun and getting rolled by them was even less fun

however i do also maintain there shouldn't be elo gains and losses for pug wins/losses. it should be a fixed value based on your highest level played (with maybe a cutoff, if you played advanced in like, season 3 with a 6-6 record and havent played in a trillion years) and then it should stay that way and be used to balance the pugs, similar to tf2pickup

i think open pugs are completely fine and i can only think of one blatant alt that ive seen playing RGL pugs and i think he is banned now anyway. i think the only reason open pugs can be bad sometimes is when theres a big disparity in skill between the players added up and creates uneven pugs, but thats better than never getting to play at all honestly. im also very glad most of the adv/inv players are playing captain pugs now, because getting raged at by higher level players was so unfun and getting rolled by them was even less fun

however i do also maintain there shouldn't be elo gains and losses for pug wins/losses. it should be a fixed value based on your highest level played (with maybe a cutoff, if you played advanced in like, season 3 with a 6-6 record and havent played in a trillion years) and then it should stay that way and be used to balance the pugs, similar to tf2pickup
221
#221
7 Frags +

did the pugs quietly change to 30 minute + progressive overtime? i remember when the new servers were added people just assumed they hadn't been updated but i definitely prefer 20 minutes if it is going to be on the progressive ruleset and cant remember the last time i played one of those

did the pugs quietly change to 30 minute + progressive overtime? i remember when the new servers were added people just assumed they hadn't been updated but i definitely prefer 20 minutes if it is going to be on the progressive ruleset and cant remember the last time i played one of those
222
#222
9 Frags +

why does this website create pugs it knows are going to be terribly unbalanced

why does this website create pugs it knows are going to be terribly unbalanced
223
#223
1 Frags +
Wild_Rumpusdid the pugs quietly change to 30 minute + progressive overtime? i remember when the new servers were added people just assumed they hadn't been updated but i definitely prefer 20 minutes if it is going to be on the progressive ruleset and cant remember the last time i played one of those

yeah wtf 0_0 -- was specing/playing the pugs today, and the timer is back to 30 minutes.

Hoping its temporary b/c I like the 20 minute timers also.

[quote=Wild_Rumpus]did the pugs quietly change to 30 minute + progressive overtime? i remember when the new servers were added people just assumed they hadn't been updated but i definitely prefer 20 minutes if it is going to be on the progressive ruleset and cant remember the last time i played one of those[/quote]


yeah wtf 0_0 -- was specing/playing the pugs today, and the timer is back to 30 minutes.

Hoping its temporary b/c I like the 20 minute timers also.
224
#224
8 Frags +
Wild_Rumpusdid the pugs quietly change to 30 minute + progressive overtime? i remember when the new servers were added people just assumed they hadn't been updated but i definitely prefer 20 minutes if it is going to be on the progressive ruleset and cant remember the last time i played one of those

I think the RGL updater plugin was overwriting the modified config files, should be fixed now hopefully

[quote=Wild_Rumpus]did the pugs quietly change to 30 minute + progressive overtime? i remember when the new servers were added people just assumed they hadn't been updated but i definitely prefer 20 minutes if it is going to be on the progressive ruleset and cant remember the last time i played one of those[/quote]
I think the RGL updater plugin was overwriting the modified config files, should be fixed now hopefully
225
#225
6 Frags +

I get the reasoning behind the first come first serve for open pugs, however if it results in favorability that looks like this, also resulting in 5-1 games giving negative elo for winning players, then it shouldn't be a thing in place. The ultimate priority for the open pugs should be to favor an even matchup/favorability. I think that if a certain player is fatkidded 2 times in a row, they get to get a guaranteed slot in the third pug, or a similar system. I understand getting fatkidded repeatedly is frustrating but I think the frustration of being placed in an unwinnable pug is way higher.

I get the reasoning behind the first come first serve for open pugs, however if it results in favorability that looks like [url=https://i.imgur.com/PGEEYbD.png]this[/url], also resulting in [url=https://i.imgur.com/NER09HA.png]5-1 games giving negative elo for winning players[/url], then it shouldn't be a thing in place. The ultimate priority for the open pugs should be to favor an even matchup/favorability. I think that if a certain player is fatkidded 2 times in a row, they get to get a guaranteed slot in the third pug, or a similar system. I understand getting fatkidded repeatedly is frustrating but I think the frustration of being placed in an unwinnable pug is way higher.
226
#226
4 Frags +

I'd be curious to see how the new elo system that's being worked on will tackle the issue of a player having different skill levels for each class

atm almost every single pug I've been picked on soldier has been poverty because my soldier has the same elo as my scout despite the skill disparity between my two classes being huge... I'd like to play demo too but I know I'll get yelled at for being ass :/

I'd be curious to see how the new elo system that's being worked on will tackle the issue of a player having different skill levels for each class

atm almost every single pug I've been picked on soldier has been poverty because my soldier has the same elo as my scout despite the skill disparity between my two classes being huge... I'd like to play demo too but I know I'll get yelled at for being ass :/
227
#227
6 Frags +

you can do first come first serve and still avoid favorability that bad. have the system get the pool of 12 ppl that will be in the pug (and even what classes they will be on, if you want to make optimization simpler) and then have find the combination of players such that the total rating is the closest to even. if you lock the classes, there should be only 32 ways to slice it which should take literally no time even if you use just brute force to find the most even favorability.

you can do first come first serve and still avoid favorability that bad. have the system get the pool of 12 ppl that will be in the pug (and even what classes they will be on, if you want to make optimization simpler) and then have find the combination of players such that the total rating is the closest to even. if you lock the classes, there should be only 32 ways to slice it which should take literally no time even if you use just brute force to find the most even favorability.
228
#228
9 Frags +

I've also been fatkidded twice in a row or fatkidded before people who came after me so I'm not sure that it's even trying to do first come first serve

I've also been fatkidded twice in a row or fatkidded before people who came after me so I'm not sure that it's even trying to do first come first serve
229
#229
17 Frags +

Can we get a pug cancel button? Just played an open pug and the other team's Demo instantly subbed himself out, unable to find a sub for 20 minute pregame (plus an additional 2 or 3 minutes because the pug doesn't automatically cancel once the server time limit ends in pregame). Which essentially meant 11 players were stuck in a pug that wasn't going to start. Most people left by the 15 minute mark, and both teams readied with 2 players on each team. And while I don't really care about my elo, it sucks having to lose some just trying to get the dead pug over, just so we can get into the next pug.

https://i.imgur.com/vBzt3uN.png

https://logs.tf/3709351
I think a cancel button that requires between 8 to 10 players to go through would be pretty reasonable to avoid situations like this from happening again.

Can we get a pug cancel button? Just played an open pug and the other team's Demo instantly subbed himself out, unable to find a sub for 20 minute pregame (plus an additional 2 or 3 minutes because the pug doesn't automatically cancel once the server time limit ends in pregame). Which essentially meant 11 players were stuck in a pug that wasn't going to start. Most people left by the 15 minute mark, and both teams readied with 2 players on each team. And while I don't really care about my elo, it sucks having to lose some just trying to get the dead pug over, just so we can get into the next pug.
[img]https://i.imgur.com/vBzt3uN.png[/img]
https://logs.tf/3709351
I think a cancel button that requires between 8 to 10 players to go through would be pretty reasonable to avoid situations like this from happening again.
230
#230
10 Frags +
tobiasCan we get a pug cancel button?

We played a pug where the enemy med did not have tf2 installed and no sub was found in 20 minutes. Enemy team had to go afk and let us win in 5 minutes to end the pug.

[quote=tobias]Can we get a pug cancel button? [/quote]

We played a pug where the enemy med did not have tf2 installed and no sub was found in 20 minutes. Enemy team had to go afk and let us win in 5 minutes to end the pug.
231
#231
0 Frags +

is it more difficult to get picked in the open queue if you don't have any match history? i've been fat kidded by system maybe 5 times.

is it more difficult to get picked in the open queue if you don't have any match history? i've been fat kidded by system maybe 5 times.
232
#232
1 Frags +

https://imgur.com/gallery/why-why-are-comebacks-discouraged-rgl-pugs-YrElinD

Holy shit man, I hope this new elo system FIXES EVERYTHING -- because this sum' bullshit.

https://imgur.com/gallery/why-why-are-comebacks-discouraged-rgl-pugs-YrElinD

Holy shit man, I hope this new elo system FIXES EVERYTHING -- because this sum' bullshit.
233
#233
2 Frags +
Juustinhttps://imgur.com/gallery/why-why-are-comebacks-discouraged-rgl-pugs-YrElinD

Holy shit man, I hope this new elo system FIXES EVERYTHING -- because this sum' bullshit.

heavily favored team, only won 5-4, what part of this doesn't make sense? I'm against having a like 5-1 win totally tank the elo, but if you almost lose a pug you shouldve (theoretically) stomped than this seems fair.

[quote=Juustin]https://imgur.com/gallery/why-why-are-comebacks-discouraged-rgl-pugs-YrElinD

Holy shit man, I hope this new elo system FIXES EVERYTHING -- because this sum' bullshit.[/quote]
heavily favored team, only won 5-4, what part of this doesn't make sense? I'm against having a like 5-1 win totally tank the elo, but if you almost lose a pug you shouldve (theoretically) stomped than this seems fair.
234
#234
9 Frags +

how is this favorability score even calculated tho? Like if the ELOs are already not very well defined (lack of players, statistics), then some algorithm telling me I should win the game by a lot and then lose ELO for winning by a little seems kinda off. I think you should always gain ELO for winning, even if it's just a bit if the favorability is in your favor or whatever. Idk if that is something that happens in other games where an ELO system is used, but in chess for example ur not gonna lose ELO for taking too many moves to beat someone who has a lower ELO than you.

how is this favorability score even calculated tho? Like if the ELOs are already not very well defined (lack of players, statistics), then some algorithm telling me I should win the game by a lot and then lose ELO for winning by a little seems kinda off. I think you should always gain ELO for winning, even if it's just a bit if the favorability is in your favor or whatever. Idk if that is something that happens in other games where an ELO system is used, but in chess for example ur not gonna lose ELO for taking too many moves to beat someone who has a lower ELO than you.
235
#235
5 Frags +
DolphInN- Winning a PUG no longer decreases your elo, regardless of other factors

https://www.teamfortress.tv/post/1094326/rgl-pugs-public-alpha
it's literally broken

[quote=DolphInN]
- Winning a PUG no longer decreases your elo, regardless of other factors
[/quote]
https://www.teamfortress.tv/post/1094326/rgl-pugs-public-alpha
it's literally broken
236
#236
2 Frags +
dbkheavily favored team, only won 5-4, what part of this doesn't make sense? I'm against having a like 5-1 win totally tank the elo, but if you almost lose a pug you shouldve (theoretically) stomped than this seems fair.

How about YOU LOSE NO ELO AT ALL.

Also, did you even look at the teams? The 'favorability' is like that because dflamez, despite having esea history + 160 logs, doesn't have any elo. And glxy, his highest 6s szn was Intermediate, but he's played up to inv highlander, still have like 1300 elo.

Realistically, the favorability should be like more even, like 0.4 - 0.5 or simply even at best.

Also, we can't talk about 'favorability', knowing it doesn't account ETF2L history, so if prem scout decides to add up, they can fuck the favorability scales.

[quote=dbk]
heavily favored team, only won 5-4, what part of this doesn't make sense? I'm against having a like 5-1 win totally tank the elo, but if you almost lose a pug you shouldve (theoretically) stomped than this seems fair.[/quote]

How about YOU LOSE NO ELO AT ALL.

Also, did you even look at the teams? The 'favorability' is like that because dflamez, despite having esea history + 160 logs, doesn't have any elo. And glxy, his highest 6s szn was Intermediate, but he's played up to inv highlander, still have like 1300 elo.

Realistically, the favorability should be like more even, like 0.4 - 0.5 or simply even at best.

Also, we can't talk about 'favorability', knowing it doesn't account ETF2L history, so if prem scout decides to add up, they can fuck the favorability scales.
237
#237
23 Frags +
dbkJuustinhttps://imgur.com/gallery/why-why-are-comebacks-discouraged-rgl-pugs-YrElinD

Holy shit man, I hope this new elo system FIXES EVERYTHING -- because this sum' bullshit.
heavily favored team, only won 5-4, what part of this doesn't make sense? I'm against having a like 5-1 win totally tank the elo, but if you almost lose a pug you shouldve (theoretically) stomped than this seems fair.

jordangod combo scout should default to giving his team zero favorability

[quote=dbk][quote=Juustin]https://imgur.com/gallery/why-why-are-comebacks-discouraged-rgl-pugs-YrElinD

Holy shit man, I hope this new elo system FIXES EVERYTHING -- because this sum' bullshit.[/quote]
heavily favored team, only won 5-4, what part of this doesn't make sense? I'm against having a like 5-1 win totally tank the elo, but if you almost lose a pug you shouldve (theoretically) stomped than this seems fair.[/quote]

jordangod combo scout should default to giving his team zero favorability
238
#238
9 Frags +

To Dolphin:

If your current directions in Elo are failing, here is a free open source library you can try to use- https://openskill.me/en/stable/manual.html#user-manual

You can read about the results on a sample overwatch dataset here: https://arxiv.org/abs/2401.05451

This is essentially one of several ports of this paper: https://jmlr.csail.mit.edu/papers/volume12/weng11a/weng11a.pdf
It's in python but there are other implementations in other languages.

To Dolphin:

If your current directions in Elo are failing, here is a free open source library you can try to use- https://openskill.me/en/stable/manual.html#user-manual

You can read about the results on a sample overwatch dataset here: https://arxiv.org/abs/2401.05451

This is essentially one of several ports of this paper: https://jmlr.csail.mit.edu/papers/volume12/weng11a/weng11a.pdf
It's in python but there are other implementations in other languages.
239
#239
13 Frags +

yo can we have a vote to draw option for pugs that make you serve a 40-year prison service

yo can we have a vote to draw option for pugs that make you serve a 40-year prison service
240
#240
17 Frags +

https://media.discordapp.net/attachments/992861299656642612/1285035718506446878/image.png?ex=66ed6b86&is=66ec1a06&hm=c01075f7772905580215f86e8b74fb6173010fe5ac0c0873245fd2994e0d866b&=&format=webp&quality=lossless

[img]https://media.discordapp.net/attachments/992861299656642612/1285035718506446878/image.png?ex=66ed6b86&is=66ec1a06&hm=c01075f7772905580215f86e8b74fb6173010fe5ac0c0873245fd2994e0d866b&=&format=webp&quality=lossless[/img]
1 ⋅⋅ 5 6 7 8 9 10
Please sign in through STEAM to post a comment.