Note 48 games across six volumes Revised twice
The Map Room emblem The Map Room Games and getting lost
Running survey

The Compass Tax

Twenty-nine games with a compass ribbon, nineteen without, all played without guides and all drawn from memory a week later. The gap between the two groups has widened every volume, and the claim has had to be narrowed twice.

48 games · six volumes · revised twice, both revisions below

A compass ribbon is the strip across the top of the screen showing bearings and nearby points of interest. It is nearly universal in large games, it is genuinely useful, and it appears to cost more than anybody accounts for. That is the claim, and this note is the evidence for it plus the two places it has already been proven too strong.

The raw split

Of the forty-eight games logged, twenty-nine shipped with a compass ribbon on by default and nineteen did not. Recall was scored by the drawing test — a sketch made a week after finishing, judged full, partial or none.

GroupGamesFull recallPartialNone
With compass ribbon2931115
Without191261

Ten per cent full recall against sixty-three per cent. That is a large enough gap that the interesting question stopped being whether it is real and became what exactly is causing it, because the obvious explanation is wrong.

First revision: it is not the compass, it is what the compass points at

The original version of this note, written in volume two, claimed the ribbon itself was the problem. Volume three broke that. Three games arrived with compass ribbons showing only cardinal bearings and no points of interest, and all three scored full recall — the same as the no-compass group.

A bearing is information about the world. A marker on a bearing is an instruction. The first supports navigation and the second replaces it, and because almost every commercial ribbon does both, the two had been conflated. The claim was narrowed accordingly: a compass costs nothing, and a compass populated with destinations costs a great deal.

Superseded, volume three: "A compass ribbon reliably degrades a player's mental map." The three cardinal-only games disproved this directly, and the sentence was wrong as written.

Second revision: density matters more than presence

Volume five forced the second change. Two games with fully populated ribbons scored partial recall rather than none, and both had the same unusual property: fewer than thirty points of interest in the entire world. With so few markers, most travel happened with an empty ribbon, and the players' own routes had to carry the load.

So the figure that actually predicts recall is not whether a ribbon exists but what share of travel occurs with a destination active on it. That is why Marker Fields at ninety-four per cent produced four dots on a page, and why the same game's final ten hours at fifty-eight per cent produced the only region I learned.

The current claim, stated carefully

Recall correlates with unguided travel share, not with interface presence. A world learned during the six per cent of travel that is unguided will be learned six per cent as well, roughly, and every navigational convenience that raises the guided share pays for itself in geography the player never acquires.

Stated that way, the compass is not the villain. The villain is destination density, and the compass is simply the most common delivery mechanism for it.

A traveller on a high ledge holding an open paper map above a valley of forested ridges
The cleanest case in the whole survey: Fathom Coast, zero guided travel, full recall, and half the size of the worst performer.

What would overturn this

One specific result would break the current claim: a game with high guided travel share and full recall. I have not logged one in forty-eight games, but I have also not logged a game that deliberately tries — a fully marked world with landmarks so strong that the geography sticks anyway.

The design that would produce it is not hard to imagine. Mark every destination, but make the routes between them so distinctive that they are remembered involuntarily. Nobody in this sample has attempted it, and until somebody does, the claim above stands as the best available description rather than a law.

Three smaller findings

Elevation visibility matters independently. Games where high ground can be seen from low ground score better regardless of interface, which is part of why Tallow Quarter's terraced city was learned so quickly.

Player-drawn maps are not exempt. Lantern Deep showed that a good self-made map substitutes for memory in exactly the way a minimap does, just more slowly and with more affection.

Recall is not comfort. The three games in the no-compass group that players would find hardest to travel all scored full recall, and one of them was actively unpleasant to navigate. Nothing in this survey should be read as a claim that unguided worlds are more enjoyable.