We all know that after the auto-tests (and before re-doing the auto-tests…) you gotta edit manually. I like to read through the code from beginning to end a few times, making corrections, but to get spacing and transitions and general feeling (and so on) right, there’s nothing like a proper play-through.
When you do a play-through (of either your game or beta-ing for someone else) do you:
*Make the choices you’d make playing the game?
*Test the choices that are most likely to have errors (eg “they/their/them” for pronouns)?
*Take all the first choices, then all the second, then all the third?
*Test one type of player (eg villainous) then another (heroic), then another (all about the romance) and so on?
Or some combination, or something totally different?
I do an initial run-through as the character I’d play if I were playing the game as a consumer. Then I do the exact opposite of what I’d play as and make all the opposite choices I’d want/took in the first playthrough! After that, I just speed through by clicking everything’s first choice…
I test my playthrough’s exactly the way I play any game that I enjoy (since I am a completionist anyway, and HAVE to read all of the choices or I will DIIIIIIIE): 1st playthrough is all of the choices/romances I would choose 1st, normally. 2nd playthrough is my 2nd choices (since there are usually hard choices or similar choices in/from my 1st playthrough) then I start doing opposites and speed throughs. Then when I think I have done it all - I go back and end with a combination all of my favorite choices/scenes from the previous playthroughs so that I close with my best possible game…I am really more laid back than I sound, I swear… hahaha
When I was trying an open beta I usually start with how I will play the game normally, then in the next playthroughs I did almost exactly the same except for some different choices so when an error popped out/something changed I noticed it immediately.
I did a run thru like this before my first private beta version, just to make sure that the game was cohesive, because I edit down the files for posting (to remove half-written scenarios). It was surprisingly fun.
I have a couple of subroutines to make setting gender variables a little easier, so now and then I change my default testing character from male to genderfluid, to assess pronoun selection.
Like I said, I have one main testing character whose personality traits are based on my understanding of the core of the character, and I take him through all three paths pretty regularly. I also have a second, less well-developed testing character. She’s not his “opposite”, but she does differ from him on some critical traits. In my debug code, I have them listed as “charming manipulator” and “straightforward intellectual” respectively.
That’s the big thing. I have a ton of debug code that lets me skip over major chunks of game, so that I can test whatever I’m working on right now. That means I have some structured paths that I take over and over, which I know is a vulnerability, but that’s why I rely on beta testers, too. Basically, I have a “quick choice” debug menu at the start of every chapter, that sets the variables to the values they would have at the end of a particular set of choices.
I have a random character creator that makes a random character, and then I just do randomtest runs and read the result as a story and see if it flows.
Write a section. Test that section. With my testing variable set to true, include a shortcut to my next section with default stats. Write the next section. Rinse and repeat.
Every so often I do a full run-through. Not often though. Right now a run through of my game can have something like 300-500 screens, I reckon. Not actually sure the exact number. It’s a lot though.
Yes. All of the above, varies depending on what stage of the process I’m in or if I’ve just made some big changes and I want to focus on a specific bit of code/story/etc. Similar to what @Szaal said, I’ll write a big section or branch, then cycle through various playthrough methods before moving on to writing the next part.
Helps to break it down in manageable chunks so I’m not just doing all the tests at once and making a huge “to fix” list for myself. That sounds really tedious and daunting!
I’m pretty much on board with @Szaal and @MizArtist33 - I obsessively playtest whatever section I’m working on, and fix things, and as soon as it works properly for a few times, I release it to the ghosts of the machine and never touch it again.
I just hope that that things make sense at the end of it all, and I read my randomtest stories, similar to what @malinryden said, to see if things flow properly.
