plum
Account Details
SteamID64 76561198145759044
SteamID3 [U:1:185493316]
SteamID32 STEAM_0:0:92746658
Country United States
Signed Up June 14, 2018
Last Posted November 17, 2024 at 10:54 PM
Posts 85 (0 per day)
Game Settings
In-game Sensitivity
Windows Sensitivity
Raw Input  
DPI
 
Resolution
 
Refresh Rate
 
Hardware Peripherals
Mouse  
Keyboard  
Mousepad  
Headphones  
Monitor  
1 2 3 4 5 6
#7 rahThread: AI players in TF2 General Discussion

I’ve thought about related topics in this field (multi-agent reinforcement learning) before, so here’s a partially thought out answer if you really care.

The most relatable result would be in 2019, when OpenAI Five beat the reigning Dota 2 TI Champions OG in a BO3. This was a big deal, since not only is Dota 2 a complex game, the AI would have had to navigate cooperation between players on the team while working adversarially against others. However, it’s kind of hard in practice to create a TF2 version of this, even if you ignore sniper/scout. The primary barrier is data/computational power. The idea is that the AI learns by playing games against itself repeatedly. The amount of computational power needed to have the AI get to a pretty high level is most likely too expensive for peons. Not only is the training expensive, it’s also not cheap to consistently run the AI and get them to play games against players. I also don’t know how convincing the Dota 2 match was since I don’t play that game (it may have suffered the same hallucinations as the early AlphaGo versions), but generally if you want to improve performance, you just need to run more iterations of self-play, which would cost even more resources. Now maybe there exist people that have access to these resources and are able to design a reinforcement learning algorithm for TF2, so it’s theoretically possible.

However, it’s quite an open problem to create an AI that “mimics” real players in these complex problems. When these AI are trained, they are aiming to win the game, not to act like a player at some level. Thus, if you are able to play against such AI as described before, I think you could learn a thing or two, but you could definitely feel that you are playing against bots and not humans.

Let me know if you have any questions!

posted 9 months ago
#6 I want your weird and unserious frags! in Videos
Brock you dont even know what the fuck I'm referring too between him and me
...
God forbid anyone is transparent on this shitshow.

When you apologize to someone about a private matter you do it privately. Hope this helps!

posted 9 months ago
#39 really bad tf2 memes (ill start)... in TF2 General Discussion

https://cdn.discordapp.com/attachments/1011306481343287329/1192286058277240993/0g2hK4eVxN5Ag3s5_RbUtM8pqFd_thhMMz_xNRl7_ANOjLblyZzvsFI74Tt1GHKlWUzEyHTPtgWxs6pDLjDZGaqZB5349Qb-lY-M.png?ex=65a885eb&is=659610eb&hm=c66b7a165069bc96b985fa8205b65f9bf9f81d20f038d188c6a9426839f7fc37&;

posted 10 months ago
#38 really bad tf2 memes (ill start)... in TF2 General Discussion

https://cdn.discordapp.com/attachments/1011306481343287329/1192286059019632821/RDT_20210916_1919101222956450134594346.jpg?ex=65a885ec&is=659610ec&hm=1ea17df00ba1201b35346248a627bae47b933445d70b472638c423de06ec544c&;

posted 10 months ago
#62 Admins Unjustly Force a Forfeit in Grand Final in TF2 General Discussion
workingondyingYeah I was just speculating there and throwing out the idea to get activity into the forum.

So you just want attention. Well congrats, you got it but at the cost of looking like a clown to the international TF2 scene.

posted 11 months ago
#61 RGL PUGs Public Alpha in Projects
alfredodanElo in team games are bad.

I just want to be clear I think the Elo rating system is good for 1v1 games like chess, but not as good in 6v6 for example. For a matchmaking system to actually work, there needs to be some skill estimation used. Valve uses a modified Glicko system for example, with MMR being a skill estimation. The key thing here is that people should not know their MMR or rating, it should only be used for the matchmaking system.

bearodactyljust let people captain lol humans are better at picking than some dumb system

I simply disagree. I feel like this only works for circlejerks where everyone knows everyone else.

posted 11 months ago
#56 RGL PUGs Public Alpha in Projects

I'm not entirely sure how the Elo is implemented here, but it seems like some initial guess is based on some past results. Ideally people play a couple pugs (kinda how you play placement games) so that their Elo is closer to their true skill. Again I'm not entirely sure about the details of implementation (Elo is not designed for team games. I have many problems with this), but I wouldn't claim the matchmaking system doesn't balance teams well based on 3 games.

posted 11 months ago
#36 really bad tf2 memes (ill start)... in TF2 General Discussion

https://preview.redd.it/payloads-better-v0-zi0rdfic39wb1.jpg?auto=webp&s=6cf3520f60d6b343cf41beb120a2d0d8113e945b

posted about a year ago
#5 where r lan pics in TF2 General Discussion

I know many people are asking me but the issue is I'm not doing the post-processing and dunno when it will be finished (I have asked). There are also some LAN videos/interviews that I hopefully can get out, but I am waiting on permission.

posted about a year ago
#1 Disturbing trend among new players in TF2 General Discussion

Ever since I got last place in newcomer in RGL S2, I've been permanently stuck in open. This has allowed me to observe a disturbing trend among new players coming into competitive TF2.

Newbies now type .ss instead of !log to see the logs, without knowing what sizzling stats even is. This is due to the following server message sent after each map finishes: "To see the stats, type .ss"

Are we (serveme) teaching these new players bad habits?

posted about a year ago
#30 Why use P-REC over integrated demo support? in TF2 General Discussion

Not sure if you care about the other direction, but I delete P-REC sometimes since it crashes valve comp.

posted about a year ago
#34 rgl discussion thread in TF2 General Discussion

Note that RGL has not responded to this thread. RGL has previously said a simple Q&A thread is not going to address everything so they did a transparency stream that has yet to be uploaded (or done again). But please make a Q&A or respond to questions here. Here are some of mine btw:

I remember RGL saying somewhere they were a for profit organization. What exactly does this mean since none of the staff get paid? On that note, is RGL ever going to get any sponsorships?

What exactly is the RGL mission statement because the old one by sigafoo sucks ass. Is it to increase the team count? Increase viewership? Maintain status quo? Focus on Prolander? It can be multiple, but having a mission is helpful because it guides decisions better than having players vote on things.

posted about a year ago
#5 Are We Back Bros? in Off Topic
hamahamHow can we transform these numbers for a push for Valve to stop ignoring the competitive scene of the game and perhaps even listening to us for feedback in case there is a Source 2 Port?

The numbers have everything to do with the summer update and nothing to do with competitive TF2. Not sure how one can make the leap to say Valve should invest anything into competitive (using these numbers). We've reached the point where the competitive community needs to grow the scene by themselves to make it worthwhile for Valve to think about us.

posted about a year ago
#4 analysis of prem/invite matches? in TF2 General Discussion

Some examples:
aim: in depth analysis

botmode:
localhost 1 vod

localhost 2 demos

marmaloo:
froyo vs g6

kaidus:
fullerton lan 1 pt 1
fullerton lan 1 pt 2

check those vod dump channels for other players/more vids

posted about a year ago
#2 playoffs clips that'll make you sad :( in Videos

another banger. thank you

posted about a year ago
1 2 3 4 5 6