Hacker Newsnew | past | comments | ask | show | jobs | submitlogin

This is pretty interesting.

I play a lot of games, but never got into fighting games.

Seeing the strategy "laid out" might be a useful way to understand the mechanics rather than figuring them out with trial and error.

Trial an error can be fun (e.g Zelda), but it's always been something that stopped me getting too far with this genre.



If you want to give it a try, Street Fighter 6 is really, really good. Aside from just being an amazingly well thought-out fighting game[1], it's also got a big player base so you'll have plenty of other noobs to play with in the ranked matchmaking mode, and the networking code is perfect[2] so you can just hop on and start having fun.

If you do decide to jump in, just remember that you will be terrible at the start. Everyone is. No, you don't have to understand frame data and have a bunch of sick combos and understand all of the mechanics to play online. Just hop in, the game will match you against players that are about at your level. SF6 is my first fighting game, I started at the very bottom of the ranked ladder (like literally, the very bottom) and now I've played more than 80 hours this year and I've just now started breaking into the "good" ranks (Platinum). In the lower ranks I had no combos and never used the super gauge at all and still won and had a great time. Don't stress too much about your rank, just think of it as a way to match you with good quality opponents.

Anyway. Can't recommend it highly enough.

[1] Here's a very in-depth overview: https://words.infil.net/w04-sf6review.html

[2] They use rollback networking, which is the gold standard for online multiplayer games https://arstechnica.com/gaming/2019/10/explaining-how-fighti...


The second article mentions delay based and rollback both of which seem to be appropriate for peer to peer graphs, but there is a third option: server authority. I created a game with mechanics very roughly analogous to fighting games (nebulous.io) where there is no client side prediction, no rollback, and a 20hz update loop. Clients just render what the server reports (with some smoothing). I never get complaints about fairness, and players can quickly learn to adapt to their latency to servers located in different regions. Its as well now as it did in 2015, when a significant portion of players used 3g or worse cellular connections.


I suggest you read again about rollback input based deterministic simulation, it is the superior option by far, especially for competitive, you have in practice “no lag”.

That you never got any complaints doesn’t means your solution doesn’t have plenty of issues, also 20hz is pretty low.

There is a reason why no multiplayer fighting will even be considered competitive if they don’t use rollback netcode in this day and age.


This?

https://play.google.com/store/apps/details?id=software.simpl...

You think this is roughly analogous to a fighting game?


Yes. Although I'm not going to quote my exact wording. Feel free to scroll up and/or familiarize yourself with the game if you're having trouble understanding.


To add to it - the frame meter in SF6 training mode is an equivalent of the cards presented in the article, but built into the game.




Consider applying for YC's Winter 2026 batch! Applications are open till Nov 10

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

Search: