Home / Blog / Audience is voting on a frame

The audience isn't judging your worth, they're voting on 15 seconds

The Omoggle audience is faster than people think. The median time a viewer takes to click a vote, measured across roughly a quarter-million server-recorded votes, is 3.1 seconds. Three-quarters of votes are in within 6 seconds. The "15 seconds" in our match window is a hard upper bound; almost nobody uses all of it.

That's the number this whole article is about. The voter who decided against you took three seconds to do it. A vote isn't a verdict. It's a click on the side of a page.

What three seconds of attention actually looks like

Willis & Todorov's 2006 paper on first-impressions showed that observers' judgments of trustworthiness, attractiveness, and aggression off a 100-millisecond exposure to a face barely changed when the exposure was extended to a full second [1]. The original finding has been replicated in variations: the snap judgment doesn't get more accurate with more time. The viewer effectively decides instantly and then rationalises.

That's not unique to Omoggle, or to face-rating. The same mechanic runs every time you swipe past a stranger on a dating app, click a thumbnail on YouTube, or scroll past a face on Instagram. It is the default mode for handling high-volume strangers. The viewer doesn't sit with you. They click and the next match loads.

What the viewer is actually weighing

Across the literature on rapid face evaluation, a few features get weighted disproportionately in snap judgments:

Eye region attention bias (40–50% of first fixations land there) [2].
Brightness/contrast — well-lit faces win in paired comparisons before any other variable.
Tiny expression cues — a micro-smile or relaxed jaw shifts perceived friendliness more than facial geometry does.
Order — the side shown first wins slightly more often, a known framing effect [3].

None of those are about your face's "intrinsic" attractiveness. They're about how the camera, the moment, and the interface rendered you in those three seconds. The audience is voting on a rendering of you.

The forgetting curve

A viewer on Omoggle queues five to twenty matches in a session on average. By the time they're three matches downstream of yours, the chance they remember anything specific about your face is, generously, very low. Ebbinghaus's 1885 forgetting-curve work is the canonical reference here [4] — most details of a brief stimulus are lost within 20 minutes if there is no rehearsal.

The practical implication, which is honestly the point of this whole article: the audience that voted against you is not still thinking about you. They didn't tell their friends. They don't remember the match. They might, in fact, be the same person who voted for you on the next match and completely failed to register that they did either.

What I tell players who write in

We get three or four emails a week from players who had a rough run. The thing I keep saying:

You're carrying the result. The audience isn't. They had a three-second decision, a click, and the next match loading. You have the memory of the loss and the rest of your day. That asymmetry is the entire reason a bad match feels so much louder than its actual weight — and it stops being loud the moment you notice the asymmetry exists.

If you want to test the "audience doesn't remember" thing

Queue ten matches, then ask yourself how many of the opposing faces you can describe. Most players, including me, remember roughly zero in any specific detail. The asymmetry runs both ways: just as your opponents don't carry you out of the room, you don't carry them. That's almost always a relief once you notice it.

A small honest counterpoint

The "they don't really see you" framing is true at the level of an individual three-second vote. It is less true across hundreds of votes — at that scale, the audience is sampling a real distribution, and your average performance does carry some real signal. If you're 200 ELO below where you want to be after 200 matches, that's information. The thing you shouldn't do is collapse 200 matches' worth of inference into the single match that just hurt.

Close the tab and come back when the result has stopped feeling like a verdict. The arena is still there.

Or queue a match →

Sources & references

Willis, J., Todorov, A. (2006). First Impressions: Making Up Your Mind After a 100-Ms Exposure to a Face. Psychological Science, 17(7), 592–598.
Hsiao, J.H. & Cottrell, G. (2008). Two fixations suffice in face recognition. Psychological Science, 19(10).
Tversky, A. & Kahneman, D. (1981). The framing of decisions and the psychology of choice. Science, 211(4481).
Ebbinghaus, H. (1885). Über das Gedächtnis (English translation: Memory: A Contribution to Experimental Psychology). Foundational forgetting-curve work.
Omoggle server-recorded vote latency dataset, 2024–2026, anonymous.