combinatorics / Random graphs; thresholds and expectation thresholds

The second Kahn-Kalai conjecture

Claims Theorem 1.1: for every n≥2n\ge2 and every finite simple graph HH with h≥1h\ge1 edges and at most nn vertices, pc(n,H)≤min⁡{1,2048e50pE(n,H)(1+log⁡2h)}p_c(n,H)\le\min\{1,2048e^{50}p_E(n,H)(1+\log_2 h)\}, hence pc(n,H)≤6144e50pE(n,H)log⁡2np_c(n,H)\le6144e^{50}p_E(n,H)\log_2 n. The logarithm depends only on the number of edges of HH, which is stronger than the conjectured log⁡n\log n. The proof builds on Tran's spread-link cores and the Mossel-Niles-Weed-Sun-Zadik resampling, with a tree reduction replacing Tran's proposed coupling conjecture. Constants are explicit but huge; it says nothing about optimal constants or about hypergraph targets.

36Significance / 100
1Frontier events
0Verification tasks
0Recorded attempts

Temporal state

Current frontier

No reconciled state yet.

Append-only history

Frontier timeline

combinatoricsSep 24, 2026Significance 36/100Registry: lean checked

The second Kahn-Kalai conjecture

Prior state unknown→proved

Claims Theorem 1.1: for every n≥2n\ge2 and every finite simple graph HH with h≥1h\ge1 edges and at most nn vertices, pc(n,H)≤min⁡{1,2048e50pE(n,H)(1+log⁡2h)}p_c(n,H)\le\min\{1,2048e^{50}p_E(n,H)(1+\log_2 h)\}, hence pc(n,H)≤6144e50pE(n,H)log⁡2np_c(n,H)\le6144e^{50}p_E(n,H)\log_2 n. The logarithm depends only on the number of edges of HH, which is stronger than the conjectured log⁡n\log n. The proof builds on Tran's spread-link cores and the Mossel-Niles-Weed-Sun-Zadik resampling,…

SourceReplayReproducedFormal proofStatement auditExternal checkExpert reviewPeer review

Research memory

Claims and attempts

Scoped claims

Source authenticated

Claims Theorem 1.1: for every n≥2n\ge2 and every finite simple graph HH with h≥1h\ge1 edges and at most nn vertices, pc(n,H)≤min⁡{1,2048e50pE(n,H)(1+log⁡2h)}p_c(n,H)\le\min\{1,2048e^{50}p_E(n,H)(1+\log_2 h)\}, hence pc(n,H)≤6144e50pE(n,H)log⁡2np_c(n,H)\le6144e^{50}p_E(n,H)\log_2 n. The logarithm depends only on the number of edges of HH, which is stronger than the conjectured log⁡n\log n. The proof builds on Tran's spread-link cores and the Mossel-Niles-Weed-Sun-Zadik resampling, with a tree reduction replacing Tran's proposed coupling conjecture. Constants are explicit but huge; it says nothing about optimal constants or about hypergraph targets.

Claims Theorem 1.1: for every n≥2n\ge2 and every finite simple graph HH with h≥1h\ge1 edges and at most nn vertices, pc(n,H)≤min⁡{1,2048e50pE(n,H)(1+log⁡2h)}p_c(n,H)\le\min\{1,2048e^{50}p_E(n,H)(1+\log_2 h)\}, hence pc(n,H)≤6144e50pE(n,H)log⁡2np_c(n,H)\le6144e^{50}p_E(n,H)\log_2 n. The logarithm depends only on the number of edges of HH, which is stronger than the conjectured log⁡n\log n. The proof builds on Tran's spread-link cores and the Mossel-Niles-Weed-Sun-Zadik resampling, with a tree reduction replacing Tran's proposed coupling conjecture. Constants are explicit but huge; it says nothing about optimal constants or about hypergraph targets.

Recorded attempts

Evidence graph

Connected research record