Getting Started with Knight Gameplay Test Part 1
I ran into this when a few people on the discord were asking about validation steps before submitting their builds. Knight Gameplay Test Part 1 is basically the initial stress test the engine runs on any knight-class character before the full compilation pipeline kicks in. It checks collision meshes, hit registration frames, and animation state transitions. You don't need anything fancy to run it, just the standard dev environment and the test suite pulled from the git repo. Most people skip straight to the full build because they're impatient. That's why their sword attacks feel floaty in the live environment. The part 1 test catches timing mismatches between your swing animation and the hitbox activation window. Without it, you're guessing. I spent three days once chasing a bug where a two-handed slam was registering one frame late on certain frame rates. Turns out the test would have flagged the offset at 60fps but passed at 30fps, which is exactly the kind of inconsistency that slips through. Navigate to the /tests/knight directory and execute the batch script with the flag enabled. The output logs go into the debug folder. Look for the section labeled state_transition_check. If you see any red lines there, those are your problems. Most beginners miss that the test needs a clean save state. Dirty profiles cause false negatives about twelve percent of the time. I delete my local save before every run now. It adds two minutes but saves me from chasing ghosts.
The biggest issue people run into is that the test assumes a baseline sensitivity profile. If you've customized your input curves, the hit registration numbers will be off. There's no warning in the readme about this, which is annoying. You either reset your cfg back to default or add a manual override parameter to the command line. I just keep a stock config file on my desktop and swap it before testing. Takes thirty seconds. Another thing nobody mentions is memory usage. The test loads the full knight model with all equipped assets, even if you're only testing basic attacks. On machines with less than sixteen gigs of RAM, it will stutter and give you unreliable timing data. I learned that the hard way on an older setup. Running it on a cloud instance fixed the issue entirely, though it added latency to my workflow.
When it doesn't help
Let me be clear about the limits here. This test only covers the knight's default moveset. If you're working on a modded weapon or a custom ability tree, Knight Gameplay Test Part 1 won't validate any of that. You'll need to run the extended suite in part 2, or write your own test case. The framework is open enough that you can add to it, but there's no official documentation for that process. You're mostly on your own once you go past the base class tests. The other real limitation is that the test results are deterministic based on the hardware clock speed. Different CPUs will produce slightly different frame timing data. If you're submitting results to a leader board or a verification queue, make sure you're running on recommended hardware. Otherwise your numbers might look wrong even though your build is fine. Download link is in the repo readme under the releases section. Grab the latest stable version and verify the checksum before extracting. I've seen people run corrupted builds because they skipped that step, and then they blame the test instead of the archive. Happened to me once too.
Get the Full Details
