My best players don't fade out, they vanish. Three months of data from my browser strategy game.

TL;DR: One of my most active players messaged me today to say he’s taking a break. He’d played every single day for 80 days. That message sent me into the database. I pulled three months of activity data from the live world of my browser strategy game: 667 players joined, 470 played exactly one day, and 48 became what I’d call invested, playing on 14 or more separate days. Twenty of those invested players have since gone quiet, and almost none of them wound down slowly. Twelve of the twenty played most days of the week right up to their final week, then stopped. The reasons cluster around one thing, a world without enough people in it, and I’ll explain why I’m not panicking about that yet.

A goodbye message started this

This arrived this morning, from the third most active player in the game. I’ve trimmed it and taken the names out.

I’ve been setting self mini challenges to keep engaged, upgrade mines to max etc, but I feel like I’m playing a solo idle game. That’s on me for the record not the game. I actually think the game is perfect, it just has end game players who don’t want to kill each other else the game dies. It just needs the people. […] Me and [another player] have debated just kamakazing on each other lol. But with no risk of islands being taken then there’s no danger and no point. […] I haven’t felt like I belonged to a game for long time and I feel it here.

He put his account into holiday mode and stepped away. The player he’d joked about kamikazing into did the same thing, the same day.

I checked his history before replying. He had a session on 80 out of 80 days since joining. Not 80 days with some gaps. Every single day for eleven and a half weeks, and he was suggesting improvements to the daily check-in screen the day before he left. You don’t get a cleaner picture of an invested player going quiet, so instead of guessing at what’s happening to my mid-game, I spent the day being a data scientist about it. Everything below comes from the live database of one world, bots and my own accounts excluded.

Most players leave on day one. That’s a different problem.

First, the honest baseline. 667 players have touched this world since it opened in late February.

70% played one day and never came back. That number sounds brutal but most of it is cold traffic behaving like cold traffic. A Reddit post in mid-April brought 311 players in one week, and 266 of them, 85%, played exactly one day. The world’s weekly actives spiked to 354 that week and glided back down ever since.

The one-day churn is a first-session problem, and it’s a known one. This post is about the other end of the chart. It’s a strategy game you check a few times a day, built to be played over weeks, so the players who matter most are the small group at the bottom of those bars.

48 players produced 59% of all the play

I’m calling a player invested once they’ve been active on 14 or more separate days. There are 48 of them, 7% of everyone who ever joined, and between them they account for 59% of every play-day the world has ever had. The whole game, socially, is four dozen people.

Of those 48, 23 are still active this week. Five have been quiet for over a week, and 20 have been gone for two weeks or more, which at this point I count as left. The ones who left didn’t leave quickly. The median departed veteran was around for 28 days, and two were around for 70. If they’d grabbed a random mobile game off the app store, most of them would have deleted it inside a week. These people gave a month, minimum two weeks, to a game that isn’t at version one yet.

They play every day right up to the end

The question I actually wanted answered: when an invested player leaves, does their activity taper off, or does it just stop? Here is every one of the 20, week by week from joining, with the number of days they played each week.

Twelve of twenty are cliffs. Look at the rows: 7, 7, 7, 6, gone. 7, 7, 6, 5, gone. One player was on six or seven days a week for ten straight weeks and then simply never logged in again. Only three of the twenty show the slow wind-down I expected to find, where the weekly days drift from 7 to 2 to 1 before the end.

That changes how I think about retention warnings. By the time a dashboard shows an invested player’s activity declining, the decision already happened. There’s almost no taper to catch. The only real window is the five players who’ve been quiet for 8 to 13 days right now, and the data says that window is about a week wide before quiet becomes gone.

The leavers were the ones talking to me

126 players, 19% of everyone who joined, have sent me messages in the game. Eleven of the twenty departed veterans were regulars in my inbox, reporting bugs, suggesting features, asking design questions. These weren’t anonymous churn rows. I knew them.

One found the game by searching for Inselkampf, the old German browser game mine descends from, something he said he does every few years to see if anyone has built one. He played 33 days out of 33, reported bugs and an exploit along the way, and then one afternoon asked a pointed question about a change that protects active players’ islands from being taken. Fifty-six minutes later he posted his exit, and his reasons were specific: the new colonising requirements, the mechanics drifting away from the classic formula, and the world events filling his screen. He wanted the game I started building, not the game it’s becoming. His last words in the thread were “thank you for the fun time.”

Another went quiet for a few days, and the game punished him for it. He came back to attacks in progress, and a later visit showed his islands mostly gone, eaten by the cleanup system that exists to recycle abandoned territory back into the world. His message that day: “Nothing, game is bullshit.” Here’s the part that surprised me though. His heaviest five days of play came right after the first attack, and even after the angry message he kept checking in, trading with another player once, most recently two days ago. Anger, it turns out, is still engagement. The system that ate his islands was built for a full world where leaving means leaving. In a small world it can’t tell a holiday from a quit.

And then there’s the player I know nothing about. Forty-seven active days across ten weeks, six or seven days a week, the most invested player to ever leave. He sent me two messages in his whole time in the game. He stopped one day in mid-May, no complaint, no goodbye, and the cleanup system removed his abandoned account a few weeks later. Half the loss looks like this, and no amount of reading my inbox will explain it.

The peace is what’s killing them

Put the goodbyes together and they say the same thing from different angles. Today’s message: “it just has end game players who don’t want to kill each other else the game dies… with no risk of islands being taken then there’s no danger and no point.” Six weeks earlier, the current number one player, still active, 161 islands, told me the mirror image: the best growth strategy would be taking islands off other players, “but I don’t really want to do that since it’s a new game without many active players, and I don’t want to cause any of them to quit.”

Sit with that pair of quotes for a second. The strongest players are deliberately playing nice to protect a small world, and that niceness is exactly what’s emptying it. This is a persistent world browser strategy game whose whole engine is risk, fleets that can be intercepted, islands that can be taken, and the people at the top have informally switched the engine off. The most invested players responded by inventing mini challenges for themselves, maxing mines for the sake of it, and when the self-assigned content ran out, they put themselves on holiday. Another active veteran asked me outright in May: how many players are actually active? I wonder why there’s not more traffic here. The thinness of the world is visible from inside it.

So the mid-game leak isn’t a content problem or a balance problem. The early game is solved enough that people reach the middle, and the late game works if you’re a warlord. The middle is where you need other people, and there aren’t enough other people. It just needs the people, like the man said.

Why I’m not panicking

I keep coming back to the same reframe. If these players had downloaded a game off the app store, would they have played it every day for a month, for eleven weeks? I download games, play them for a few days, and delete them. By that standard, a game that isn’t at version one holding people for 28, 47, 80 days isn’t haemorrhaging, it’s overperforming with the wrong population size.

The plan doesn’t change. Get to version one, then push for the 100 daily players where the dynamics change, where there are enough people that an attack has someone behind it and a defence has stakes. There’s a forum and a subreddit where this game hasn’t been properly posted yet. Posting now, before it’s ready, would pour more cold traffic onto a world that can’t yet hold the warm players it has, and it would leak more of them. Losing invested players before v1 stings, but every one of these exits is a marked leak I’ll know how to watch when it matters.

And the part I didn’t expect to be the headline: people write goodbye letters to this game. They put themselves in holiday mode instead of deleting. They say “thank you for all your hard work” on the way out and tell me they’ll rejoin when the world fills up. Twenty years of downloading and silently abandoning games, and I can’t remember thanking a single one. Something here is worth coming back for. It just needs the people.