Hi everyone, I’m a computer science graduate student, casual league player, and AI enthusiast, and I want to take a minute to explain why I think a strong AI for League would be enormous for the game both for the casual player and for esports orgs.
First, let’s address the feasibility of a League-playing AI: fairly recently, OpenAI (https://openai.com/five/) created a team of AI players that defeated world champions at Dota2. Dota and League are similar enough that their methodology/setup should be able to learn League as well as it learned Dota, if given access to the League backend (Riot pls). Obviously, this would be a lot of work, so now let’s explore why that work would be worth it.
Humans learn from stronger competition. One great example of how stronger computers improve human gameplay can be found in chess, where players at lower levels benefit from having good opponents regardless of where they are or who they know, and masters benefit from objective analysis of their moves when reviewing their games. Right now in League, the best humans have to experiment to discover the optimal strategies each patch; often, this takes a very long time. Since there are relatively few elite human players, things like “ardent censor is the most busted item in the game” can take months to discover. AI teams could play many more games at the Challenger/Professional level than human Challenger/Pros can, which could add statistical significance to winrates. Strong AI could inform pros (and the balance team) of strong picks very quickly. Getting a good “read” on the meta would be less based on luck.Загрузка...
Better practice for everyone: part of why it’s hard to climb is that you have to break bad habits that you form against weaker opponents that can’t punish you. Having the chance to lane against a strong bot to practice matchups, CSing, wave control, or any other mechanics, could make it much easier to intentionally improve, even for causal players.
Regional parity: China and Korea have long enjoyed the advantage that their pro teams get to scrim each other, and huge player bases make it easier to find high quality, high Elo solo queue games in EUW/KR/China. A strong AI team could allow pro teams to A) scrim against an opponent that is always trying and doesn’t tilt, and B) hide those strategies from other teams. This benefits top teams by giving them more quality opponents, but it also benefits lower tier teams that often can’t get scrim blocks with the top teams in their region.
Sustainability for pros: because of the size of the game, pros need to put in huge numbers of games to learn important matchups. Currently, they also have to spend time searching for strong counterpicks to meta matchups, play them, and also convince analysts or coaching staff that the matchup is as good as they think. Having a strong AI explore promising matchups for them (and even being able to watch how the AI plays the lane/map in those matchups) could decrease the number of games outside scrims that pros need to grind. This would decrease injury risks and help prevent burnout.
In conclusion, a strong League-playing AI could offer great benefits to players at every level of the game, and especially to pro players and orgs. I really hope Riot will look into supporting development for an AI for League.
Source: Original link
© Post "Why League of Legends needs a strong AI" for game League of Legends.
Top 10 Most Anticipated Video Games of 2020
2020 will have something to satisfy classic and modern gamers alike. To be eligible for the list, the game must be confirmed for 2020, or there should be good reason to expect its release in that year. Therefore, upcoming games with a mere announcement and no discernible release date will not be included.
Top 15 NEW Games of 2020 [FIRST HALF]
2020 has a ton to look forward to...in the video gaming world. Here are fifteen games we're looking forward to in the first half of 2020.