Finding the gap nobody is looking at

Some numbers are over-picked because they're memorable. Those numbers will hurt you. The gaps next to them will help you.

Cognitive science calls them salient numbers: numbers that stand out in human memory and pull attention regardless of whether they're a good strategic choice. They're the worst picks in a Lowball round because too many people gravitate to them, and your pick collides with theirs.

The good news: if you can list which numbers are salient, you can list which numbers are not, and the not-list is where you should be picking.

The salient numbers

Number	Why it's salient
1	The smallest possible win. Every level-1 reasoner picks it.
7	Cross-cultural "lucky" number. Studies of random-number picking find 7 is picked at ~3x its base rate.
13	Counterintuitively over-picked — people pick it because they think nobody else will. This is also true of 13.
17	Asked to "pick a random number between 1 and 20", a famously disproportionate share of people pick 17. Try this on friends.
10, 20, 25, 50, 100	Round numbers feel "memorable" — they double as anchors. Anything ending in 0 or 5 is over-picked.
42	If you've heard of Douglas Adams, you know.
69, 420, 666	Internet humor numbers. Skip them unless you're confident your opponents are very serious.

That's not a complete list, but it covers most of the high-collision numbers I've watched in real round data.

The gaps

For each salient number, there's a negative-space band right next to it that's unusually empty. Examples:

• 2, 3, 4 are crowded (people dodging 1). 5 and 6 are crowded (people dodging 1 and 7). 8 and 9 are the first real gap above the low cluster.
• 11 and 12 are the gap above 7. 14, 15, 16 are the gap below 17.
• 9 and 11 are the gap around 10. 24 and 26 are the gap around 25.
• 22, 23, 27, 28, 33, 34, 36, 37: numbers that don't look like anything and therefore get under-picked.

The simplest version of the gap strategy: pick a number that's not a single digit, not a multiple of 5, not 7 or 13 or 17, and not ending in 0. That filter alone eliminates 80% of the heavy-pick zone.

Off-by-one is the most reliable single move

If 25 is going to attract 20 picks today, then 24 and 26 will each attract dramatically fewer. The off-by-one numbers are the cleanest negative space because the human brain rounds. People who think "around 25" pick 25, not 24 or 26.

This works in both directions. Around 10 → 9 or 11. Around 50 → 49 or 51. Around 100 → 99 or 101 (assuming the round's ceiling is high enough).

What the Stats panel will show you

After each round closes, its picks join the Stats panel: a histogram of every number picked, the most-chosen and most-winning numbers, and the middle 50% of picks — filterable by Yesterday / Week / Month / Ever. (The open round stays hidden, so nobody can peek mid-game.) Three ways to use it:

• The most-chosen number tells you exactly where the biggest collision happened. That number, and its close neighbors, are crowded territory.
• The middle-50% range tells you where half the crowd plays. Numbers just outside that band — especially below it — are the thinnest, most interesting ground.
• Compare Yesterday against the Month view. A drifting distribution means the crowd is adapting; a sticky one means habits are doing the picking.

One thing not to do

Don't pick the same number every day. If you become predictable, you'll get punished on the day you pick a number that the rest of the crowd has also figured out is "your" number. Variety in your own picks is part of staying in the negative space.

Mix it up. Today pick 19. Tomorrow pick 8. The day after, pick 37. Watch what wins. Adjust.

---

More reading: Level-k thinking, applied to Lowball · Streak tactics