Slow Playthrough Quarterly emblem Slow Playthrough Quarterly Games played to the end
Issue 14 · Essay three Four games finished 413 recorded hours
Essay · Issue 14

Endings We Did Not Want: On Finishing Games That Should Not Finish

The most common failure in a long game is not a bad ending but an ending at all. Four hundred hours of accumulated habit is interrupted by twenty minutes of resolution, and the resolution is asked to carry a weight nothing that short could bear.

Read 19 min413 hours across four gamesAll finished
An empty wooden rowing boat moored at a weathered jetty on a still lake at dusk with a lit cabin window in the distance
Every long game I have finished ends, in memory, at a jetty. I do not think this is a coincidence.

I have finished four very long games in the past two years, totalling four hundred and thirteen recorded hours, and in each case my strongest memory of the ending is of the fortnight before it. Not the conclusion. The delay.

In three of the four, I stopped within sight of the end and did other things for between nine and twenty-three days. I tidied inventories. I revisited places that no longer contained anything I needed. In one of them I spent an entire evening walking a route I had walked perhaps sixty times, for no reason I could articulate, knowing that the final sequence was one loading screen away.

This behaviour is common enough that it must be telling us something about what these games are, and I do not think what it tells us is flattering to the endings.

The arithmetic of resolution

Consider the proportions. A four-hundred-hour game concludes in a sequence lasting, in my four cases, between eighteen and forty-one minutes. That is between one-fifteenth and one-thirtieth of one per cent of the experience, and it is asked to do the work that in a novel is distributed across a final quarter.

GameHours loggedEnding lengthDays delayed
First14334 min23
Second9618 min9
Third11241 min16
Fourth6227 min0

The fourth is the interesting one. It is the shortest, and the only one I finished without hesitating, and it is also the only one whose ending I remember with any warmth. I do not think the ending was better written. I think a sixty-two-hour game has not yet had time to become a place, and so leaving it is an event rather than a departure.

The longer a game succeeds at being somewhere you live, the more completely its ending must fail. This is not a design flaw. It is a structural condition of the form.

Three responses, none satisfactory

Designers know about this problem and have three standard answers to it.

The first is to return you afterwards: the ending resolves, and then you are placed back in the world with the final sequence pending indefinitely. This is the most common answer and the most incoherent one, because it purchases continuity at the cost of meaning. Nothing that can be un-happened carries weight.

The second is to make the ending enormous — a final act of twenty or thirty hours that gradually closes systems down. Two of my four attempt this. It works better than it sounds, and it has the peculiar effect of making the last twenty hours the most attentive, because for the first time in months the game is going somewhere.

The third is to refuse to end, which is not really an answer but is honest about the problem. I have a great deal of sympathy for it, and no interest in playing anything that adopts it, which tells you something about my own inconsistency here.

A dusty wooden shelf holding a stack of unlabelled cartridge cases beside a small brass key in warm window light
Files I have not deleted from games I finished years ago. There is no mechanical reason to keep them.

What the fortnight was for

I have come to think the delay is not avoidance but a form of reading. During those nine to twenty-three days I was not putting off the ending. I was conducting one, privately, at a length proportionate to the time I had spent — revisiting, tidying, walking the route. The game's ending is eighteen minutes long. Mine was three weeks, and it was better.

If that is right, then the criticism to make of these endings is not that they are weak but that they are redundant. The player has already been ending the game for a fortnight. What arrives afterwards is a formality that interrupts something more complete than itself.

A modest proposal

I would like one of these games to notice. Not to add a system — the moment this becomes a mechanic it dies — but simply to be aware that a player who spends nine days doing nothing in particular near the end is doing something, and to be as unhurried about it as the previous four hundred hours were.

The fourth game on my table did something close to this by accident. In its final week, the ambient dialogue of its settlements changed slightly: nothing dramatic, just people mentioning that things had been quiet lately. I noticed on my third evening of aimlessness, and it was the only moment in four games where I felt the ending had been built for somebody who had actually lived there.

Then it ended in twenty-seven minutes, and I sat for a while, and I did not delete the file.

Recorded

413 hours

Four long games finished across two years, with delays of twenty-three, sixteen, nine and zero days. No score is given, here or anywhere else in this journal.