00:02 <+bridge> [ddnet] ty @deen c: 00:03 <+bridge> [ddnet] xd what happend? closed lul 00:04 <+bridge> [ddnet] rip me getting in the web dev list of ddnet 00:06 <+bridge> [ddnet] i like how the yt hash starts with ``eZ_`` xd 00:22 <+bridge> [ddnet] "Oh, didn't see that, made the same change right now. Thanks anyway." 00:22 <+bridge> [ddnet] https://github.com/ddnet/ddnet-web/pull/41#issuecomment-376690076 01:20 <+bridge> [ddnet] ye i found that too 01:20 <+bridge> [ddnet] so he was doing the same thing xd and found my pr aftwards? 01:20 <+bridge> [ddnet] thats sad .-. 01:21 <+bridge> [ddnet] how was he even that fast 01:21 <+bridge> [ddnet] the video was public 20 secs 02:58 < jxsl13> :O 07:58 <+bridge> [ddnet] @ChillerDragon Just happened to be ingame where the video was announced to be uploading/uploaded 07:58 <+bridge> [ddnet] So I checked it out immediatel. Only checked my emails afterwards, that's when I saw your PR 10:08 <+bridge> [ddnet] who want's a free VPS xD 10:19 <+bridge> [ddnet] @deen nice that you are fast at updating videos at frontpage .-. 10:19 <+bridge> [ddnet] Sad that i didnt make it into the web contributor list xd i really want to get in there 10:19 <+bridge> [ddnet] @ReiTW | Blizzard yes sir 12:18 <+bridge> [ddnet] @ChillerDragon You'll find something useful to contribute, don't worry. Or just run a spell checker over it.. 12:20 <+bridge> [ddnet] lot's of things to fix on that sit 12:22 <+bridge> [ddnet] lots of things to fix on that site 12:44 <+bridge> [ddnet] @Learath2 do you have some example? c: 12:48 <+bridge> [ddnet] the ridiculously long server specific ranks pages (ranks/solo), the misaligned "Points calculation" part on ranks page 13:09 <+bridge> [ddnet] oh true thats missaligned 13:23 <+bridge> [ddnet] But those might not be in ddnet-web repo 13:40 <+bridge> [ddnet] i suck at aligning things anyway xd 13:52 <+bridge> [ddnet] uhm does some1 know how i extend my bunch of .acc files by another value? 13:52 <+bridge> [ddnet] 13:52 <+bridge> [ddnet] i tried 13:52 <+bridge> [ddnet] ``` echo "0" >> *.acc ``` 13:52 <+bridge> [ddnet] 13:52 <+bridge> [ddnet] but it didnt let me .-. 13:52 <+bridge> [ddnet] 13:52 <+bridge> [ddnet] ``` -bash: *.acc: ambiguous redirect ``` 13:52 <+bridge> [ddnet] `for file in *.acc; do echo "0" >> "$file"; done` 13:53 <+bridge> [ddnet] uff thats a lot of typing 13:53 <+bridge> [ddnet] shit bash 13:53 <+bridge> [ddnet] but ty c: 13:58 <+bridge> [ddnet] BTW, has anyone tried to teach a neural network to play TW in a fasion similar to OpenAI Gym environments? https://gym.openai.com/ 13:58 <+bridge> [ddnet] yes 13:59 <+bridge> [ddnet] idk his name anymore was some dude trainign it on gores 13:59 <+bridge> [ddnet] he was live on twitch 13:59 <+bridge> [ddnet] it was very interesting to watch o.O 13:59 <+bridge> [ddnet] I myself haven't seen neural nets playing TW yet – but a couple of other AI strategies 🙂 13:59 <@deen> @bob_fasul If you do so for DDRace with 2 players, please call it Siamese twin AI 14:00 <+bridge> [ddnet] 2p neural net o.O 14:00 <+bridge> [ddnet] it would be so great 14:00 <+bridge> [ddnet] I actually meant https://blog.openai.com/universe/ xD 14:00 <@deen> U 14:00 <+bridge> [ddnet] I actually meant xD 14:00 <+bridge> [ddnet] well, it would be nice to have multiple agents cooperating indeed 14:00 <+bridge> [ddnet] but I doubt it would be easy to teach it DDRace 14:01 <+bridge> [ddnet] especially since some maps require lots of thinking (e.g. Koule 7, where 3 players must cooperate from the very first stage) 14:02 <@deen> just train it on teehistorian data :D 14:02 <@deen> lots of successful teams playing 14:02 <+bridge> [ddnet] yea, might also work, first some supervised learning on teehistorian or just random demos, and then unsupervised learning on various novice maps 14:03 <+bridge> [ddnet] bet it would be very good at solo maps : 14:03 <@deen> Well, the first problem would be how to define progress, how to get closer to finish line 14:03 <+bridge> [ddnet] just train it on @N9mkOik's races 14:04 <+bridge> [ddnet] 😛 14:04 <+bridge> [ddnet] (supervised learning is when the correct answers are known beforehand, unsupervised is when it tries to guess the answers based on some raward) 14:04 <+bridge> [ddnet] (supervised learning is when the correct answers are known beforehand, unsupervised is when it tries to guess the answers based on some reward) 14:05 <+bridge> [ddnet] well, we might get some people to finish the race, and then guess progress based on position compared to position of those people 14:06 <+bridge> [ddnet] I mean, get the nearest position on their path, then check what was their time at that point -- closer to finish would be closer to end time 14:08 <+bridge> [ddnet] @heinrich5991 might work, why not 14:49 <+bridge> [ddnet] @ChillerDragon what's a web contributor list? 15:00 <+bridge> [ddnet] https://github.com/ddnet/ddnet-web/graphs/contributors @Rockus.~ the only ranking what matters 15:04 <+bridge> [ddnet] ah okay 19:31 <+bridge> [ddnet] why is there still no fix for players with high times (300min for example) listed at the bottom of the scoreboard 19:31 <+bridge> [ddnet] just increase the default value from -9999 to -999999 20:18 <+eeeee> or the players could just hurry up and finish faster 23:16 <+bridge> [ddnet] no u