Hacker News new | past | comments | ask | show | jobs | submit login

If you think DotA2's matchmaking is unfriendly to new players, I don't think you ever experienced DotA1 matches. Every single game was wildly unbalanced, and as a result of one or so players, every single game was an absolute stomp where the losing team never had a chance.

The skill gap between good and bad players is ludicrous. It's like putting LeBron James against middle schoolers. There simply isn't a fun game when the game is determined by the singular best player, and the other nine players are being carried to victory regardless of their contribution, or being pounded into the dirt because they're less than exceptional.

I could never go back to random matchmaking after experiencing global rankings.




Of course, using a bad metric is often better than using no metric at all, but it doesn't follow that the current strategy is optimal.

It especially doesn't follow that global rankings should be a user-visible metric, or an explicit player motivator (a la seasonal rankings). People don't bot because they're unhappy with the players they're matched against. They bot because they want a publicly visible number to get bigger and because player rewards are locked behind competitive rankings.


But there are no bots available that can play the game for you competently. Even deepmind had to have quite a few restrictions to do so and such a system most likely can't even learn an evolving metagame.

People who join ranked want to compete. If you don't want to then there is another queue available that's not a public ranking.


> But there are no bots available that can play the game for you competently.

Then who cares whether or not players bot? Let them. Heck, go all the way and actively support it if it's not an issue.

On the design side, I'm seeing a number of people commenting on here that getting rid of a publicly-facing global ladder would mean that these games couldn't be competitive. I don't believe that's true. We have wildly competitive games in the physical space that don't use global ladders, and they don't suffer for it.

Global ladders are a very specific, very narrow mechanic -- they have some advantages for competitive play, and a whole lot of downsides. And we have options. At the very least, even if we do nothing else, we can shrink the size and make them regional. Or because things are digital we can throw out geography entirely and base them off clans.

Competition in small groups is often preferable to competition in large groups anyway, since that can foster rivalries, and because repeat matchups between competitive players are usually more interesting than random ones.




Join us for AI Startup School this June 16-17 in San Francisco!

Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: