Red Alert ELO-rating list

Announcements and discussion about community-run events.
noobmapmaker
Posts: 1086
Joined: Wed Dec 10, 2014 11:59 am

Red Alert ELO-rating list

Post by noobmapmaker » Sat Jul 22, 2017 12:04 pm

Hey everyone,

Had some nightshifts with nothing to do, so I decided to download ELO-software (link for download) and input all RAGL S02 (+qualifier matches), RAGL S03 (+qualifier matches) and DoomHammers 1v1 Tournament games. Alltogether ~500 games!

Note: I've only added games that where posted in black font (played games). I decided not to use scores of players that have forfeited the season or dropped out because of strikes. Simply because SoScared edited out those scores, I think. Also didnt input scores that have been won because of failed contact and thus have not really been played.

A little introduction: the ratingsystem that is used is the USCF system, used for the American Chess league. Here's a PDF with an elobarate explanation. Alot can be said about the reliability of ratings and which system to use, but I decided to use the USCF-system because the software was userfriendly and in general I think it can be used for Red Alert 1v1's just as well as for chess. As for a proper show of skill: the more games that are entered into the system, the more accurate it becomes... more or less.

I will keep updating the ratings!

Whenever there is a serious 1v1 event/tournament I will enter all played games. Also you can have your casual Online Multiplayer (1v1) games entered into this ratinglist. If you haved played a game against a player and the both of you decide it to be a rated game, then both of you comment "Rated game" in the first 30 seconds of the game and upload the game to Gamereplays.org. Also post the result to this topic as a comment. If you do all this then I will enter the game. Please note: a maximum of 3 consecutive games per person is allowed. Wait before at least one new game is entered before you can again submit rated games.

The list as is on 22th of July 2017

Initial playerrating 1300.
Link to a google spreadsheet with additional information, made possible by Netnazgul!

Image

I think the asterisk [*] next to a name means that it's still a provisional rating (played less than 25 games) and that means the rating is more volatile than those who have a more established rating - not sure though!

So if you want to enter a rated game (with a maximum of 3 consecutive games) to the list:
- decide together with your opponent before the game that the game is rated
- both players have to comment "Rated game" in the first 30 seconds
- upload your game to gamereplays.org
- comment the result in this thread

PS: if you want to see the file with all games, download the list here. You will have to install Elorater through the provided link above and then manually place it in the folder. Please note that once a game is entered it cannot be deleted! (you'll have to replace the .elo file)
Last edited by noobmapmaker on Tue Jul 25, 2017 10:20 am, edited 3 times in total.
Playlist with ALL games of the Dark Tournament Youtube.com/CorrodeCasts
Consider supporting OpenRA by setting a bounty or by donating for a server

User avatar
Clockwork
Posts: 314
Joined: Fri Jan 06, 2017 3:50 am
Contact:

Post by Clockwork » Sat Jul 22, 2017 3:55 pm

Second best player in RA ill take that 😊

User avatar
WhoCares
Posts: 287
Joined: Mon Feb 13, 2017 11:28 pm

Post by WhoCares » Sat Jul 22, 2017 4:17 pm

That's kinda nice, let's see how players will dare to have casual game rated ^^

noobmapmaker
Posts: 1086
Joined: Wed Dec 10, 2014 11:59 am

Post by noobmapmaker » Sat Jul 22, 2017 4:38 pm

I hope people will submit their games! At this time alot of games have been played in tiny habitats: Fahrrad for example has only played mastergames and lost (almost) all of them and thus is rated very low. If he'd play some minions/recruits then his rating will quickly rise to a more accurate level.

By the way: your opponent of casual rated games does not have to be on this list! If you've played a rated game against someone who is not on this list he is simply added with a 1300 starting rating.
Playlist with ALL games of the Dark Tournament Youtube.com/CorrodeCasts
Consider supporting OpenRA by setting a bounty or by donating for a server

User avatar
JOo
Posts: 510
Joined: Wed Jan 25, 2012 8:12 pm

Post by JOo » Sat Jul 22, 2017 7:16 pm

people could just abuse that by having a friend rename himself everytime and let him play like shit, play on a server that doesnt show off the IP (not even that, a lot of players just have dynamic IP's)

you could also just easily get a spectator in to spec-snipe on a different machine or via friend (similar to stream-snipe) and get advantage over the enemy,

in public 1v1's theres often a lot of spectators ... they could easily just compromise a "rated game" by calling out someones strategy or moves ...

also, do you allow even custom maps ? ... what maps are allowed ?

here something that is related

https://github.com/OpenRA/elobot

we used to have an "Elobot" on #OpenRA ... like 3-4 years ago , but people didnt really use it

Blackened
Posts: 308
Joined: Sat May 21, 2016 6:27 pm

Post by Blackened » Sat Jul 22, 2017 8:23 pm

mfw I have a higher ELO than FiveAces :lol:

microbit has a point about the potential for abuse but I think further expansion of the rules could eliminate most of it.

Still this is really cool NMM! You're killin it with all the things your coming out with! :)

User avatar
WhoCares
Posts: 287
Joined: Mon Feb 13, 2017 11:28 pm

Post by WhoCares » Sat Jul 22, 2017 9:30 pm

I think the opposite, for what i know of the competitive 1V1 community. I beleive that people are not dumb enought to try to abuse in an obvious way a ranking system in a little group of people who know each other.

Nobody would beleive me ranked second or first place in 2 weeks from now if I pile up casual matchups vs random unknown players and i would lose any credibility in the eyes of other players. "Hey, look i'm first in official ranking but everybody knows I'm a joke" (Yeah Lorry, we all think that ! ;))

Until the community becomes much more larger, his size is its better protection from that kind of nonsense.

For myself, i'll try to play as mutch as my online game as "official" when it comes to play the people who are in that list. I find interesting to tweak my own ranking to see more accuratly where I belong. And it could force me to try harder when i play better people to avoid piling up too many defeats in a row.

And to complete, if people right now are not satisfied with their ranking, they just have to challenge people above them they beleive they can beat easly to quickly reach the ranking that suits them.

noobmapmaker
Posts: 1086
Joined: Wed Dec 10, 2014 11:59 am

Post by noobmapmaker » Sun Jul 23, 2017 5:10 am

Both Microbit and WhoCares say sensible things. Yes there is certainly potential for abuse and it will be impossible to eliminate all ways to abuse it. But since the admission of rated games is out in the open (placed on forum and the games posted to gamereplays) we are able to keep an eye on it.

A couple extra rules to create some more reliability:

- This ranking is for shits 'n giggles & braggingrights. It's not 'official' in any way, no prizemoney will be awarded and I won't use this to seed tournaments. Also ratings are always an estimate and can be at best 'somewhat accurate'.
- I will update the rating every two weeks (gather the played games and enter them in order of submission).
- New players will only be added to the list if they have played at least two different opponents from the current list in those two weeks, or if they participate in an official event, like RAGL. If the new player has played just 1 opponent, then the games are not accepted as rated games.
- Existing players may play a maximum of two new players per two weeks, and no more than 3 games per new player.
Playlist with ALL games of the Dark Tournament Youtube.com/CorrodeCasts
Consider supporting OpenRA by setting a bounty or by donating for a server

User avatar
anjew
Posts: 522
Joined: Sat Nov 08, 2014 4:16 am

Post by anjew » Sun Jul 23, 2017 8:18 am

JOo wrote: here something that is related

https://github.com/OpenRA/elobot

we used to have an "Elobot" on #OpenRA ... like 3-4 years ago , but people didnt really use it
Is it still active?
Do you know how to use it?
I have have found no documentation on it

Im really interested in the automated system
Image

potato
Posts: 1
Joined: Sun Jul 23, 2017 8:40 am

potato = Stepexxx

Post by potato » Sun Jul 23, 2017 8:44 am

Good job! Just a note : Stepexxx and potato is one person.

User avatar
netnazgul
Posts: 455
Joined: Mon Jul 10, 2017 7:32 am
Location: Minsk
Contact:

Post by netnazgul » Sun Jul 23, 2017 10:16 am

Would be glad to see this as an ingame feature.

About abuses: as it is open like nnm said, community can ban certain games I suppose. Or boo certain players and request their rating reset.

[lots of text deleted about wishes on implementing ranked games into engine which is obviosly lots of time investment and not be seen in a forseeable future]

On another notice: can you make this a text list instead of picture? Because it's quite hard to find anything in a pictured text :P
noobmapmaker wrote: Alot can be said about the reliability of ratings and which system to use, but I decided to use the USCF-system because the software was userfriendly and in general I think it can be used for Red Alert 1v1's just as well as for chess.
Main problem of Elo rating usually is rating inflation, when there are a lot of comers/goers that just drop some rating points to the pool and leave; those who stay in the system longer will then achieve ridiculously high ratings. Is there a k-factor reduction for top players in your calculations?
Another problem of this rating system is when player achieves some high-placed ranking but then goes AWOL for a long period of time. He is not easy to overcome (because he doesn't lose when he's away) and if he returns, his ranking position would not show his actual skill.

PS Why is 1300 being a default? I thought 1000 is usually taken as a start value...

PPS
noobmapmaker wrote: Alot
:lol: http://hyperboleandahalf.blogspot.com.b ... thing.html

noobmapmaker
Posts: 1086
Joined: Wed Dec 10, 2014 11:59 am

Post by noobmapmaker » Sun Jul 23, 2017 12:10 pm

I will remember 'alot' in the future! :D not my native language but always want to learn better English :)

I believe USCF has the K-factor! Still every system, especially with limited players and data, will have it's shortcomings. It will always be just an indication of skill relative to the others in the pool. The startingvalue differs per system, this one used 1300 as default so I took that. And yeah, ingame would be nice, but difficult to implement and has been discussed often before - don't hold your breath!

Unfortunatly the software is not userfriendly when it comes to exporting the data. Maybe someone can extract it from the .elo file from the OP? It was mainly userfriendly for entering a lot of data. I can't delete games or players for instance and Ive tried to select the list but couldn't so I screencapped it.
Playlist with ALL games of the Dark Tournament Youtube.com/CorrodeCasts
Consider supporting OpenRA by setting a bounty or by donating for a server

lucassss
Posts: 144
Joined: Mon Jan 04, 2016 1:55 pm

Post by lucassss » Sun Jul 23, 2017 2:11 pm

People in the league fight mostly against people in their own division. Therefore I would guess these scores are only accurate within a division, aren't they? This seems like a really problematic method, since it give the person who is 4th place a higher ELO than the guy who is 7th in masters.

Also, no way in hell that I'm better than abcdefg30

noobmapmaker
Posts: 1086
Joined: Wed Dec 10, 2014 11:59 am

Post by noobmapmaker » Sun Jul 23, 2017 2:57 pm

We need to mingle! :)
Playlist with ALL games of the Dark Tournament Youtube.com/CorrodeCasts
Consider supporting OpenRA by setting a bounty or by donating for a server

User avatar
JOo
Posts: 510
Joined: Wed Jan 25, 2012 8:12 pm

Post by JOo » Sun Jul 23, 2017 3:33 pm

anjew wrote:
JOo wrote: here something that is related

https://github.com/OpenRA/elobot

we used to have an "Elobot" on #OpenRA ... like 3-4 years ago , but people didnt really use it
Is it still active?
Do you know how to use it?
I have have found no documentation on it

Im really interested in the automated system
currently not active , but someone could ask to activate him on the global-channel (or fork and let him join)

in order to let him join a specific server/port/channel you probably have to change line 328
on elobot.py,

https://github.com/OpenRA/elobot/blob/master/elobot.py

Code: Select all

reactor.connectTCP('irc.freenode.org', 6667, BotFactory('#openra'))
to

Code: Select all

reactor.connectTCP('irc.openra.net', 6667, BotFactory('#lobby'))
but better ask pchote, baxxster or jaZz for permission before having the bot join


.draw .help .win .outgoing .cancel .loss .incoming .verify .top .register .reject

<user> .register
<elobot> Thanks for registering, user. Your initial rating is 1200.

<user> .win vs user2
<user> .loss vs user2
(when user2 is not registered yet) -> <elobot> Can't find player "user2"

(otherwise) -> <elobot> Added unverified claim 0; ratings will be adjusted when the other party verifies it

you could then :

<user2> .verify 0 (0 is the game-ID ... which are just numbers)
(or)
<user2> .reject 0

<user> .top (ranks the top 10 players)

<elobot> #1: Scott` (1217)
<elobot> #2: Holloweye (1216)
<elobot> #3: MicroBit (1212)
<elobot> #4: pav256 (1201)
<elobot> #5: Hammster (1200)
<elobot> #6: hamb (1200)
<elobot> #7: orabot (1200)
<elobot> #8: microbit (1200)
<elobot> #9: xanax` (1200)
<elobot> #10: pchote (1200)
Last edited by JOo on Mon Jul 24, 2017 9:13 am, edited 5 times in total.

Post Reply