I Built Her a Game. I Finished It Too Late.

She would run out of lives and sit there watching the phone, waiting for it to let her play again. I was a mobile app engineer. It seemed like a problem I could solve.

The chair

She had run out of lives. So she sat and watched the phone, waiting for the timer.

Lori played Candy Crush. Match three of a thing, they disappear, you get another go. Millions of people play it. She liked it, she was reasonably good at it, and by that stage of the disease it was one of the few things she could still do on her own.

The game gives you five lives. When you lose them, you wait — half an hour a life — or you pay. That is the business model, and it is aimed at impatient people with disposable income and somewhere else to be.

Lori would run out, and then she would sit in her chair holding the phone, watching a countdown, waiting to be allowed to play again.

I would come into the room and find her doing it.

She was waiting for permission. Some product manager, somewhere, had built a mechanism to convert impatience into revenue, and it was metering out the remaining time in which my wife could still enjoy something.

What I did about it

I was a mobile app engineer. I decided to build her one that would never stop.

It is the only useful thing I could think of, and it happened to be precisely the thing I knew how to do. Twenty-odd years of writing software for a living, and here was a problem I could actually do something about.

A match-three game. Same idea, same satisfying collapse of tiles. No lives. No timer. No countdown. No purchase, no advertisement, no account, no password. You open it and you play, and you keep playing for as long as you want to, and then you stop.

I worked on it in the evenings. It took several months, which for one person building alone alongside everything else that was happening is not slow.

The disease was faster.

By the time it worked, she couldn't play it

I finished it, and she had lost the ability to use it.

Not gradually, from my side. From my side it was simply that at some point I had a working game and no player.

I never published it. It sat on a drive. It is still there.

For a long time I filed that under things I failed at. I have come to think that is the wrong shelf.

Every caregiver reading this has lost this exact race. The grab bar installed the week after the fall. The GPS tracker that arrives and she has stopped walking. The paperwork finished the month after capacity went. The stairlift, the shower chair, the sign on the door, the appointment you finally got.

You are always solving the problem she had ninety days ago. It is not because you are slow. It is because the thing you are chasing keeps moving, and it does not wait for procurement, or for surgery schedules, or for evenings after work.

That is the part nobody writes down, so I am writing it down. If you have built something excellent for a person who no longer needs it, you have not failed. You were running the same race as the rest of us.

Why commercial games stop working

They are designed to get harder. A declining player is locked out by design.

Now the useful part — because I have spent a long time since thinking about what I would build if I did it again.

Casual games escalate difficulty. That is the entire craft: keep the player at the edge of their ability so they stay engaged. It works beautifully on a player who is improving.

Applied to a player who is declining, it is a countdown to exclusion. Level forty is harder than level four, and she is not the person she was at level four. Eventually the game becomes a machine for demonstrating loss.

The rest of the standard toolkit is no better in this context:

• Lives and energy systems stop play for someone whose good hour is now, not in ninety minutes.
• In-app purchases and upsell prompts are aimed at a person who may not be able to evaluate them. This is the same vulnerability that supplement scams exploit, running inside a free game on their phone.
• Advertisements, including some genuinely predatory ones, served to someone who cannot assess them.
• Accounts and passwords that will be forgotten, locking them out of a thing they own.

None of this is malice. It is a product built for a different person.

What a game for dementia would actually need

It would have to get easier as the player got worse.

The research literature agrees, in slightly drier language. A 2024 concept analysis in JMIR Aging lists adaptability — technology adjusting to changing capabilities across the stages of dementia — as one of the core requirements for usability in this population, alongside simplicity, intuitiveness and personalization. The same work notes that as people move from mild to moderate and severe, use shifts toward doing things with a caregiver, and eventually toward experiences that require nothing of the person at all.

So the design target is not a good game. It is a game that degrades gracefully.

Fewer colors. Larger tiles. A smaller board. Then no way to lose. Then something closer to a toy than a game — a tap that produces a pleasing result, with no objective attached. Then something you do together.

No commercial studio will ever build that, because no commercial studio wants a player who is getting worse. Every metric in that industry points the other way.

If you are choosing something today, the specifics that matter — and these come straight out of the accessibility research — are:

• Large touch targets with visible borders, so it is obvious what can be pressed.
• Text labels on every icon. Abstract symbols stop meaning anything.
• No timers, no countdowns, no pressure mechanics.
• Nothing that moves, flashes or advances on its own. Auto-playing motion is strongly associated with people abandoning the task.
• Immediate feedback on every action, and a visible trace of what just happened.
• Nothing that relies on remembering something from a previous screen.
• No CAPTCHAs, no memory tests, no password-only entry.
• No congratulating them like a child. No "well done!", no "sweetie", no gold stars. This is an adult who knows what is happening to her.

When games stop working altogether

Step down early, and step down further than feels necessary.

The instinct when someone starts struggling is to encourage them — try again, you nearly had it. It is well meant and it is usually wrong. Frustration at a task that used to be easy is its own injury.

Make it easier before they ask. Fewer pieces. Bigger pieces. Remove the way to lose.

And when a game stops working at all, move toward things with no objective:

• Sorting — buttons, coins, cards, socks, a bucket of pecans.
• Folding — towels, napkins, anything that can be unfolded again later.
• Tending — watering, wiping, polishing, sweeping.
• Arranging — flowers, photographs, cutlery.
• Music. Almost always music, and almost always from when they were young.

These are not lesser activities. They are occupied hands and a sense of usefulness, which is most of what anybody wants from an afternoon. See meaningful activities in dementia for more.

What I'd do now

The same thing, sooner, and simpler.

I would not build for the person she was that month. I would build for the person she would be in a year, and let it meet her earlier version on the way through.

And I would ship something rough in two weeks rather than something good in six months. A crude game she could have used for a year would have been worth more than a polished one she never opened.

That is the lesson, and it generalizes well beyond software. The good-enough thing today beats the right thing in the spring. Whatever you are planning for them — do the cheap, ugly, immediate version of it this week.

I went back and finished it

It is a real game now, and it is free.

The code sat on that drive for a long time. I decided to finish it.

It is called Day to Day Match 3. It is the thing I described above — no lives, no timers, no ads, no purchases, no account, and no way to lose. If the board runs out of moves it reshuffles itself. There is no game-over screen in it anywhere. It gets gentler rather than harder, because that was the whole point.

Lori never played it. That part does not change, and I am not going to pretend otherwise. But somebody is sitting in a chair right now holding a phone, watching a countdown, and that I can do something about.

I started it for her and finished it for you. Day to Day Match 3 →