What a week!

Oh boy, it has been a WEEK. A week and a half.

On Monday, I launched Word Play on Steam. And the response was amazing.

At one point that first day the game showed up on the Steam home page in four different places at once! (New and Trending, Top Sellers, Games under £7, and the massive Featured & Recommended carousel). The game was also played by big streamer Northernlion.

It has also done really well in terms of sales. It surpassed the total sales for Mind Over Magnet in less than 72 hours. And right now it's sitting on just under 25,000 copies sold. I'm so thrilled. It also has a very positive rating on Steam with 96% recommended.

So, you'd think I'd be able to spend the week celebrating, resting, relaxing, and chatting to people about the game, right?

Nope!

As I've come to learn, the first week of a game's launch is just solely about fixing issues. Solving bugs, putting out updates, taking on feedback, and so on.

Now of course a huge amount of this was covered during the Steam Next Fest demo, and my QA team helped me find and fix so many bugs before launch! But... you can't cover everything.

For one thing, a game like this - with so many elements that can combine and conflict - will lead to weird edge cases that will never come up in weeks of QA testing. It's very different to a super linear game like MOM. This game has trillions of possible combinations.

And also - you've got people with super specific hardware and other variables that aren't covered by your QA team! One bug was caused by someone's (untouched) flight simulator joystick interfering with the controls!

Luckily I had built a good system for getting bug reports and feedback during the Next Fest demo.

• I pointed everyone to either my email inbox or the Discord server, so all the stuff I needed was limited to just two places.

• I put "petition this word to be in the dictionary" into the game itself, to avoid hundreds of emails like "the game wouldn't let me spell PIKACHU"

• I put lots of details into the game logs, so I could see what the game was doing when it went wrong

Each day this week I woke up, made a list of all the bugs and problems people were having, and then triaged them based on importance. From "this breaks the game and a lot of people are seeing it" down to "this is a minor visual glitch that appears rarely". That helped me prioritise.

Unfortunately, it's a famous truism of game dev that when you fix one bug, you'll likely cause another one. And as I'm still not a very confident coder, I ended up making a few mistakes that led to fresh bugs. That meant I was also fixing bugs that were caused by fixing bugs! Argh!

My biggest mistake though... oh, and this is so embarrassing.

I wanted to put out a quick hotfix to solve a small bug I had caused. I thought I had only changed one simple thing so there was no way it would cause other issues. I felt confident that I could just push this onto Steam without checking the build first.

But I had forgotten... I was also testing something else and had ticked a box that, basically, tricked a computer into thinking it was a Steamdeck. Which meant that every single person who played the game on a PC or Mac... loaded up without a mouse cursor.

Then I jumped onto the phone with a friend for a chat and didn't notice the barrage of messages and emails about it, for about half an hour! Eeep! I then glanced at my inbox and had to quickly slam the phone down, untick that box, upload a new version, and reply to everyone with an apology! What an embarrassing rookie mistake!

I can only thank that I didn't do this before going out or going to bed. That would have been a disaster.

But I'm gonna cut myself some slack. As my second only game launch I'm still learning stuff. And while this whole experience has been awesome, it has also been overwhelming. I feel so guilty when someone encounters a bug or mistake in the game. Like I've let them down. Which has led to be being overzealous with trying to get fixes out as fast as possible. Which leads to mistakes.

Time to slow it down.

And in fact, I can now slow it down. You see, I was aware of several "category one" problems with the game. You know: big, serious, game-breaking issues that were cropping up for multiple people. And for each one I could diagnose and fix it.

But one problem just kept eluding me.

It only happened on the Steamdeck, and it only happened for some people (and not me - so I couldn't replicate it). But it was crazy - the game would seem to add a random number of tiles to the bag, at weird times. You start Word Play with 80 tiles in the bag, and might add a few more as the game goes on (there's a difficult achievement for getting to 150). But these Steamdeck users were getting 200, 300 tiles in their bag by round 4. What the heck!

I racked my brains trying to think of a cause. Was the Steamdeck not powerful enough, and strange things were happening as it lagged? Was it to do with the user's system language? Was it to do with the Proton layer that converts Windows code to Linux? Nope, nope, nope.

Thankfully, one of the users who had this bug was generous enough to share the player logs with me. And patiently waited as I added more debug lines to the code (so I could see when things were firing). I found the method that was being naughty... but that was called from 18 different places in the code! So that meant more debug lines to help plot out the problem.

Only to finally figure out - aha! It wasn't a Steamdeck problem at all. It was a controller problem. I was always testing the Steamdeck with touchscreen so I didn't run into it. But as soon as I figured it out I could replicate it on Steamdeck and on my PC/Mac with a controller. In one single circumstance when playing with the controller the game didn't just move a tile around the screen but also added a copy of it to the bag.

Fixing it meant changing a single word - from true to false.

I almost cried when I finally figured it out. Partly from happiness, partly from sheer exhaustion. This has been a very intense week of highs and lows, of excitement and frustration, of success and failure.

So - time for a break. Today I've got my QA team testing this new update (version 1.05) to make sure I haven't made any new bugs (see, I'm learning!). Then when that's out and I'm happy with it, I'm taking the weekend off to stop thinking about words and golden tiles and debug logs and Steampipe, and to just play Donkey Kong, see family, touch grass, and sleep.

Thanks for the support, speak soon

Mark