Spring RTS Engine

For a really good example... Road to war <3 is 28x28, but it's mostly flat, so pathing takes very little time.

The only exception is the typemap on the road, that kinda gums it up a little.

Smoth and jK, I think you guys had some miscomm there, chill out guys.

Ok, so this is an experimental feature that in the dev builds? I guess I'll ask trepan for some examples or look on doxygen for the docs. Thanks.

Polycount does matter. It's not a major spring bottleneck, but that's because very few spring projects acctually push the limits. Several maps that have chosen to go a little feature crazy have demonstrated that FPS drops if you ignore polycount.

Mainly, my tests with PURE have indicated that the real bottleneck is how many Units get produced within a typical game timeframe. I typically play 3 vs. 1 with AIs, to test basic balance and make sure my game design is working, and I've found that when total numbers of Units climbs high enough, the costs soar rapidly. It's basically a fine balance between exponential gameplay and allowing players to have the armies in the sizes they want and expect- tough issue, but solvable.

Logically, wouldn't the AI's also be contributing additional load as their unit count rose?

Of course, and there you run into all sorts of problems, keeping track of stuff. I've done some tests, though, where I set up huge patrols on one "side", then used the other side to pick a fight, and the results were pretty similar, although less drastic.

To put it another way: if the game feels good and stable with 3AIs vs. a human player who has zero compunctions about sending massive hordes to their doom, just to keep the bad guys away from my base and economy until I'm good and ready... then it's probably going to play really well with people. That's the hope.