Click a node or edge for details

Verification — outgoing probability per source state (sums to 100%)

sourceout totalΣ% → targets
start100100%breakthrough 73.0% (73)   stuck 17.0% (17)   improved 10.0% (10)
improved39100%stuck 38.5% (15)   regressed 17.9% (7)   breakthrough 12.8% (5)   improved 10.3% (4)   end_limit 10.3% (4)   broke 5.1% (2)   churn 2.6% (1)   end_submit 2.6% (1)
churn4100%broke 50.0% (2)   breakthrough 25.0% (1)   churn 25.0% (1)
stuck99100%stuck 52.5% (52)   improved 17.2% (17)   breakthrough 12.1% (12)   end_submit 7.1% (7)   broke 5.1% (5)   end_limit 4.0% (4)   regressed 2.0% (2)
breakthrough98100%stay 55.1% (54)   end_submit 40.8% (40)   regressed 3.1% (3)   broke 1.0% (1)
stay259100%stay 79.2% (205)   end_submit 13.1% (34)   broke 3.5% (9)   end_limit 3.5% (9)   regressed 0.8% (2)
regressed15100%stuck 40.0% (6)   improved 26.7% (4)   churn 13.3% (2)   breakthrough 6.7% (1)   regressed 6.7% (1)   end_limit 6.7% (1)
broke19100%stuck 47.4% (9)   breakthrough 31.6% (6)   improved 21.1% (4)

Gap label distribution — for each label, which src→tgt edges carry it

gap labeltotal src→tgt breakdown
cosmetic_refactor102stay→stay 74.5% (76)   breakthrough→stay 12.7% (13)   improved→stuck 4.9% (5)   stuck→stuck 4.9% (5)   regressed→stuck 1.0% (1)   broke→breakthrough 1.0% (1)   stay→regressed 1.0% (1)
debug_failure16stuck→stuck 31.2% (5)   stay→stay 25.0% (4)   broke→stuck 12.5% (2)   stuck→broke 6.2% (1)   improved→stuck 6.2% (1)   regressed→stuck 6.2% (1)   improved→end_limit 6.2% (1)   churn→churn 6.2% (1)
edge_case8stay→stay 62.5% (5)   breakthrough→stay 25.0% (2)   regressed→churn 12.5% (1)
execute_plan12stuck→improved 41.7% (5)   stuck→breakthrough 25.0% (3)   stay→stay 8.3% (1)   breakthrough→stay 8.3% (1)   broke→stuck 8.3% (1)   broke→breakthrough 8.3% (1)
explore51stuck→stuck 25.5% (13)   start→stuck 25.5% (13)   stay→broke 13.7% (7)   broke→stuck 9.8% (5)   stay→stay 7.8% (4)   stuck→broke 5.9% (3)   improved→broke 3.9% (2)   start→improved 2.0% (1)   improved→stuck 2.0% (1)   stuck→improved 2.0% (1)   breakthrough→broke 2.0% (1)
fix_bug43stay→stay 20.9% (9)   breakthrough→stay 11.6% (5)   stuck→stuck 9.3% (4)   stuck→breakthrough 9.3% (4)   improved→regressed 9.3% (4)   stuck→improved 7.0% (3)   improved→improved 7.0% (3)   improved→stuck 4.7% (2)   improved→breakthrough 4.7% (2)   broke→improved 2.3% (1)   improved→churn 2.3% (1)   regressed→stuck 2.3% (1)   broke→breakthrough 2.3% (1)   regressed→breakthrough 2.3% (1)   regressed→churn 2.3% (1)   regressed→improved 2.3% (1)
implement161start→breakthrough 44.7% (72)   stay→stay 21.7% (35)   breakthrough→stay 7.5% (12)   start→improved 5.6% (9)   stuck→stuck 3.7% (6)   stuck→improved 3.1% (5)   improved→stuck 2.5% (4)   start→stuck 2.5% (4)   breakthrough→regressed 1.9% (3)   stuck→breakthrough 1.2% (2)   regressed→improved 1.2% (2)   stuck→regressed 1.2% (2)   broke→improved 0.6% (1)   regressed→stuck 0.6% (1)   broke→breakthrough 0.6% (1)   improved→breakthrough 0.6% (1)   improved→regressed 0.6% (1)
infra_confusion4stay→stay 50.0% (2)   stay→end_submit 25.0% (1)   stay→broke 25.0% (1)
optimize_perf66stay→stay 59.1% (39)   breakthrough→stay 25.8% (17)   stuck→breakthrough 4.5% (3)   improved→stuck 1.5% (1)   stuck→stuck 1.5% (1)   broke→stuck 1.5% (1)   regressed→stuck 1.5% (1)   improved→improved 1.5% (1)   improved→breakthrough 1.5% (1)   stay→regressed 1.5% (1)
revert21stay→stay 52.4% (11)   stuck→stuck 38.1% (8)   improved→stuck 4.8% (1)   churn→breakthrough 4.8% (1)
rewrite26stuck→stuck 26.9% (7)   stuck→improved 11.5% (3)   breakthrough→stay 7.7% (2)   broke→improved 7.7% (2)   churn→broke 7.7% (2)   broke→breakthrough 7.7% (2)   improved→regressed 7.7% (2)   stuck→broke 3.8% (1)   regressed→stuck 3.8% (1)   stay→broke 3.8% (1)   improved→breakthrough 3.8% (1)   regressed→regressed 3.8% (1)   regressed→improved 3.8% (1)
verify120breakthrough→end_submit 33.3% (40)   stay→end_submit 25.8% (31)   stay→stay 15.8% (19)   stay→end_limit 8.3% (10)   stuck→end_submit 5.0% (6)   stuck→end_limit 4.2% (5)   stuck→stuck 2.5% (3)   improved→end_limit 2.5% (3)   breakthrough→stay 0.8% (1)   regressed→end_limit 0.8% (1)   improved→end_submit 0.8% (1)