A difference of even 2-4f in a fighting game can make all the world of difference, both in balance and in the simple feel of the game. Games have to be designed

A difference of even 2-4f in a fighting game can make all the world of difference, both in balance and in the simple feel of the game. Games have to be designed for an expected amount of delay, which itself is often a design component of how they expect the game to function online (i.e. type and implementation of netcode). So what did we see with that genre on Stadia? Well, Stadia not only adds different input/response latency for both players in a fighting game, but it also doesn't colocate the matches to the same server in between players (or same running game instance); so the matches have the standard lag between Stadia data centers on top of the user <-> stadia delay. Unsatisfying for both players, and inferior to alternative ways of playing those games. Ironically, player-led initiatives for that kind of hosting - to deal with poor netcode of older fighting games - have shown that the very idea is sound, and can deliver a better experience e.g. Parsec game tournaments. Not optimal, but better. Google has simply failed to deliver. Part of the problem is up to the developers to solve, but the buck stops with Google. Stadia needs to deliver a world class experience on its merits, and it needs to give developers the tools and guidance and working examples - if not real games - showing how to do so. And it needed to be doing that on Day -1. Or they can just say "Stadia is for genres X, Y and Z", and only spend their time and money courting games and developers in those genres, and everyone involved would probably be better off for it. They could have - and still can - set grand but realistic expectations and deliver on them.