DO YOU PRETTY MUCH TEST AS YOU GO OR AFTER A BUILD UP OF DEVELOPMENT?

Test as I go, as often as possible. I throw in a few move events or a text pause or a sound effect or something into a cutscene and I watch that over and over again to make sure it's in the exact spot that I want it to be.

Probably I spend much more time testing than developing.....
In scenes I always want to have perfect timings for moveroutes, animations, etc.
In gameplay I always look if my technique is still working as intended etc. etc.
It's time consuming... but more testing is never bad :D

I'm sometimes a test-as-I-go person. The most likely thing that I will want to test as I write it is a sequence that has a lot of move-routes in them. Those can always mess up so bad if you're off by a step!

I probably do more testing than all other parts of development combined. I do a lot of combat testing for fine-tuning the balance and difficulty rather than just to see if it's working. This means making an enemy, testing it, changing its stats and AI, testing it, changing its stats again, testing it with two different party/equipment setups, changing its stats again, testing it with five different party/equipment setups, changing its stats again, testing it with three of those five setups a second time to make sure I got it right. And then later testing the entire dungeon as a whole and tweaking things as I go in the process, and then later testing the entire game as a whole and tweaking things as I go in the process, and then repeating those dungeon and game long tests each time I make a major change.

I think for Born Under the Rain, a four hour long game which I spent a month and a half developing, I probably spent about sixty to seventy hours test-playing it. For Vindication, a thirty hour long game which I spent years developing, I'm sure I've done well over a thousand hours of testing, maybe even more than two thousand hours.

Test as I go. I have a dedicated testing save file combined with Yanfly's Debug Extension script that I use to warp around the world and test the exact feature that was just added.

I'll also test in bulk every so often and play the game from start to finish, but this has become increasingly a problem as the game length grows.

I pretty much set myself a period of time to test and another period of time to develop/translate, so yeah, either test as I go or the testing stage will be slightly behind the development/translation stage.

If there's a point where you're not testing, you should test at that point.

Even if it's not actually playable because there's a lot of stuff you haven't made yet, you should still test with dummy characters or dummy maps or dummy equipment or dummy animations or dummy encounters or even a dummy battle system.

When coding from scratch, it can be important to test often; if you find a mistake, it will help to have the knowledge of how you coded the mistake-area fresh in your mind.