Lemma 5.2 —
ProvedBunkbedFalse.gadget_conn_allIn the gadget graph (spokes – open with probability , path edges – open with probability ), for and ,
That is: the probability that , and all lie in one component. The proof splits on which of the two spokes and are open, yielding the recurrence with and .
Formalization Note This is the partition probability for the triple , , , i.e. the probability that the percolation places all three in a single component.
import Definitions.Def_BunkbedHypergraph open Bunkbed Bunkbed.Hyper SimpleGraph Finset
namespace BunkbedFalse
/-- **Lemma 5.2** (p. 8). `ℙ_P(a ↔ v₁ ↔ vₙ) = (1-P^{2n})/(1+P)² + n(1-P)P^{2n-1}/(1+P)`. -/
theorem gadget_conn_all (n : ℕ) (P : ℚ) (hn : 3 ≤ n) (hP0 : 0 < P) (hP1 : P < 1) :
Pabc (gadgetE n) (gadgetW n P) 0 1 (Fin.last n) = (1 - P ^ (2 * n)) / (1 + P) ^ 2 + n * (1 - P) * P ^ (2 * n - 1) / (1 + P) := by
sorry
end BunkbedFalseRead-back
What the Lean code literally says, in plain math · claude-opus-5
Read-back — BunkbedFalse.gadget_conn_all
What the statement says
Fix a natural number and a rational number . The hypotheses are exactly three: , , and . Under these, the theorem asserts a single equation between two rational numbers.
The graph. The ambient vertex set is . The edge set is the set of unordered pairs arising from ordered pairs satisfying
where the second clause is read inside , so it forces . Concretely consists of
- the spokes for , and
- the path edges for ,
i.e. a fan: a hub joined to every vertex of the path . Total . (Verified by evaluation: with ; , .)
The weights. is if and otherwise. So every spoke carries and every path edge carries . Note this is a function on all of , but only its values on ever matter.
The probability quantity. unfolds, through probOf and weight, to a finite explicit sum over all subsets of :
where means " and lie in the same connected component of the graph on the full vertex set whose edge set is exactly " (SimpleGraph.Reachable, which is reflexive, so always). This is Bernoulli bond percolation: each edge is retained independently with probability , and the weights are genuine probabilities precisely because puts both and in ; outside that range the sum is still defined but is only a formal polynomial in .
The theorem instantiates this at , , . Because reachability is an equivalence relation, the event " and " is equivalent to ", and all lie in one connected component of the open subgraph". It does not say that this component is exactly — other vertices may be in it — and it says nothing about the remaining vertices. Since , the three marked vertices , , are genuinely distinct, and there are at least two path edges.
The claimed closed form. The assertion is
Here in the numerator of the second term is the natural number cast to , and is truncated natural-number subtraction. Because there is no truncation (), so the exponent means what it looks like. The first exponent is not the number of edges; the second exponent is. Division is Lean's total rational division, so both denominators would silently be interpreted as at ; excludes that.
Numerical verification
I built a scratch file evaluating both sides exactly over , and independently re-derived the left-hand side by brute-force enumeration in Python (transcribing the definitions by hand). The two implementations agree digit for digit at every shared point, which cross-validates the reading of the definitions above.
Inside the hypotheses — the identity HOLDS at every point tested. Selected values (LHS = RHS exactly):
| 3 | |||
| 4 | |||
| 5 | |||
| 6 |
Also HOLDS at for (e.g. : both sides ; : both sides ), and — by the Python enumeration — for all at those same seven values of . No failure anywhere in the hypothesised region.
Where it stops holding (both cases are excluded by the hypotheses).
- : FAILS — LHS , RHS at . This is precisely the truncation case: in , so becomes and the second term is killed by the factor while the first becomes ; meanwhile and both collapse to in , making the event vacuously true. rules this out.
- : FAILS — LHS for (i.e. ), RHS by the division-by-zero convention. rules this out.
Hypothesis strength / degeneracy notes
- appears to be unused. The identity HOLDS at (both sides , for ), at (both sides for , for , for ) and at (, , ).
- is stronger than needed. The identity HOLDS at (both sides ), at (both sides , , for ) and at (, , ). Empirically the equation is a rational-function identity in valid for every ; the only thing the two -hypotheses are load-bearing for is excluding (and giving the left side its probabilistic meaning).
- is stronger than needed. The identity HOLDS for (both sides at , at — note that here and coincide, so the second conjunct is vacuous) and for (both sides , , at ). Only genuinely needs excluding, so would suffice.
Other observations
- The proof body is
sorry; nothing is proved, only stated. - The file imports
Definitions.Def_BunkbedHypergraphand opensBunkbed.Hyper, but the statement uses nothing from that file — onlyPabc,gadgetE,gadgetWfromDefinitions/Def_BunkbedPercolation.lean(reached transitively). TheFinsetandSimpleGraphopens are likewise not exercised by the statement text. Pabchas no requirement that , , be distinct; distinctness here comes only from making , , different elements of .- The event is the "all three together" cell only in the sense of same component;
Pabcis one of five sibling definitions in the bundle (Pa_bc,Pab_c,Pac_b,Pa_b_c), none of which is invoked here.
Relevant paths: /Users/michael/Documents/misc/prove2me_workspace/Definitions/Def_BunkbedPercolation.lean (lines 33–44 for weight/probOf, 83–84 for Pabc, 109–115 for gadgetE/gadgetW), /Users/michael/Documents/misc/prove2me_workspace/Definitions/Def_BunkbedHypergraph.lean. The scratch verification file Solutions/AuditL52.lean has been deleted.
Confirmed by the mission captain (proposal self-audit).