apprehensivebob Aren't gitgud and underground crypt counter examples to that? The gameplay certainty would be a lot different without perfs, and I don't think they have a RNG problem at the leaderboard level.
Gitgud is definitely an example and not a counter example, where maps are hard enough that you can't straight up combo the whole map, that is why TP WR is faster than PRO.
Undergroundcrypt is also a hard map, and it requires a whole 10-15 minutes of basically nothing but RNG to makes it more skill based than luck based, because no one gets constantly lucky for 10 minutes nonstop on bhops. I was faster than WR on the hard crouchtunnel bhop section because I got ridiculously lucky and hit a ton of perfs in a row, but there's a whole map before and after that and I couldn't just constantly luck my way into a WR.
In return you have to learn when and when to not crouch perf on maps which can be a skill by itself. In KZT you don't do runs with crouchperfs route in mind unless it's short and easy enough that you have to put crouchperfs in the route to be faster.
apprehensivebob can't make long bhop section with scary forced perf at the end
I see that as a very good thing, if anything a forced perf at the end of a 15s long bhop section is exactly what I do not want in a KZ map. You can be really good just to lose your WR pace run because you got unlucky once, and losing 15s can easily be a run killer for any map that isn't longer than 5 minutes, which is basically 85% of KZ maps.
I don't see that as a big problem, and I don't even think that's necessarily the case. Yes you have one optimal path, but I can't do it because I don't have the strafes/techniques required to execute it, so I have to use other routes. Having skill as a barrier for routes is much better than luck in my opinion. Furthermore, hard maps will force you on one route through the bhop section anyway, regardless of mode.
And my vision for KZ bhop isn't as deterministic like SKZ is anyway, so this is a non issue in my mind.
apprehensivebob Also for mappers unconcerned about RNG then no problem, the map does not need to be competitive per comments by Nuke and Nooral.
If KZ were to have a skill based rank system using points from maps, then maps must be competitive enough in nature. And ideally RNG shouldn't be involved in determining someone's skill. Mappers are free to make their maps filled with RNG mechanics and non competitive, but those maps won't be ranked.
apprehensivebob I think it's feasible. Compared to surf, routing in KZ is a lot more separable. This should allow for solution using existing optimal control solvers (e.g. PSOPT) in a reasonableish amount of time (days for a map like colors). Certainty plenty of optimizations, complications, and work to do if relevant. I'll note I don't think every single map could be reasonably solved this way, due to having bespoke tech (e.g. some stuff with water) or too long of an uninterrupted trajectory.
I barely know anything on this, but I feel like if Valve struggled to make their own bots navigating through the map at all, I don't feel like it's very feasible, even more so without access to the dynamic system to even start finding a solution (the engine and movement code) like Valve does. Though to be fair, I doubt they tried really hard on it.
apprehensivebob There's already vastly different gameplay between maps and it's not a problem. You get maps with different block types (timed blocks, single or multi, etc), there's maps like kz_rise which introduce brand new mechanics, forced pro run maps, etc. It can be a bit confusing on some existing maps, but no real problem and it's all mappers discretion. Of course a change in RNG would need to be indicated to the player somehow, as existing maps do with their varied mechanics. While it's fine for maps to behave differently, it would be a problem if player input behaved differently on different maps (prestrafe).
The gameplay might be different but they should all be based on the same base mechanics. Every competitive de_ map uses the same guns, the same movement, the same objective, and it should be the same for KZ for the base movement code. Mappers can add whatever gimmick they want into their map (like the ability to jump off the map in vertigo), but it will be on them to port/implement it into their map, and again, there won't be any guarantee those maps will be accepted into the ranked pool.
My main gripe with KZT perfs stay: KZT perfs makes you so ridiculously fast that it makes almost every other tech irrelevant. Once people get good enough and runs become more optimized, KZT is just the race to the bottom where you spam the restart button to eventually get luckier than the previous guy.