The Bots Have Filed Their Report. Apparently We're Not Done.

Well, I guess we weren't quite done.

I announced a gameplay milestone, sent the bots off to test it, and they came back with notes.

Quite a few notes.

On the bright side, that's exactly why we built them. On the other hand, I'd briefly enjoyed the idea that the remaining list might fit on something smaller than a scroll.

Fifty to seventy-five hours. According to the robots.

The balance tests played through Swordoku using various strategies. If the reports are right, we're looking at roughly 50–75 hours of gameplay.

That's a test-based estimate, rather than a stopwatch reading for every possible human player, but I'm happy with the amount of game it suggests is there.

The reports also picked up some interesting progression patterns. Much of the equipment found during play eventually got replaced by gear bought from the vendor.

I'm fine with that. Found gear is there to provide smaller upgrades along the way, while the vendor tends to carry higher-quality equipment—for a price.

Apparently the merchant has understood his role in the economy quite well.

Too easy here. Absolutely horrible over there.

The tests identified plenty of places where the difficulty was too low or too high, so we've been making adjustments.

I want the game to be hard, particularly from dungeon 60 onward. But there's a difference between a proper challenge and discovering that a particular stretch has developed an unreasonable personal grudge.

Trying different strategies through the tests has been useful for finding those rough spots. The bots have given me much better information to work from than simply playing the early dungeons again and deciding everything seems fine.

We installed analytics. Then uninstalled analytics.

I also implemented an analytics package.

Then we deleted it.

Then we built our own system and connected it to my sales dashboard. Because apparently making a game wasn't quite enough software for one project.

The endpoints have been tested and are working, and the game is now uploading anonymous gameplay analytics to the server. The reporting pages are still being built, but the data side is coming together.

The consent handling has had attention too: opting out stops the analytics uploads, and install tracking now waits for the consent decision rather than running ahead of it.

I'm working to make those controls behave properly as part of the wider privacy and consent setup.

Once the reports are finished, they should give me a much clearer view of how people progress through the game—and where things might need another look after testing meets actual players.

And then I remembered the icons. Again.

All our equipment still needs its icons added.

Doh.

You'd think that after remembering ability icons, status icons, and all the other icons, I might have developed some sort of immunity to this particular surprise.

Apparently not.

I've also built out the achievement list, but it isn't implemented yet. There's a fair amount of tracking logic to write so the game knows when you've actually met each achievement's conditions.

Which has me considering another addition to analytics: tracking how many players reach those milestones. That could be useful for understanding progression, but it's still something I'm weighing up.

So that's two more jobs that were waiting patiently for me to remember they existed.

A little fanfare, and a few more checks

The audio list has acquired a few more entries too. When a stage ends, I need to fade the music and play a proper completion hit. Winning should sound like you've achieved something, not like the game has simply run out of monsters.

The menus definitely need sound feedback, including a distinct sound for claiming an achievement. If you've earned a reward, the button should bring a little enthusiasm to the occasion.

I also need to test the bosses' summons properly. Giving a boss the ability to call for help is one thing. Making sure the help actually turns up is another rather important detail.

Now for the office phone

On actual hardware, it's playing really nicely on my Fold 8 Ultra. Next up is the old S21 in the office, to see how it behaves on the older phone I have available.

I don't have anything older to test on, so that's the next useful check. Looking good on my main phone is encouraging, but it doesn't tell me how it'll run on everything else.

So yes, still some work to go. Audio cues, summon checks, another phone, and whatever else the list remembers before I do.

The list remains undefeated

The tests went well in the useful sense: they found things. We've got difficulty changes underway, working analytics uploads, report pages in progress, and some very visible gaps in the icon and achievement departments.

It's all moving forward. “Nearly done” just continues to contain a remarkable number of individual tasks.

The dev life never ends.