1 viewer
Price Lattice
One price per asset, fitted over the whole book, and the decomposition lattice above it filled with everything that fit could not explain. The covector is what the market agrees on; the residuals are where it does not.
- Assets fitted
- 2
- Evidence edges
- 1
- Gauge components
- 1
- Arcs
- 1
- Drawn nodes
- 1
- Bundle premiums
- 0
live price fit: covector, residual field, bundle premiums
- positions
- 1
- assets
- 2
- arcs
- 1
- drawn nodes
- 1
- premiums
- 0
- fit
- converged
Price Covector
2 assets over 1 gauge component(s), 1 pooled evidence edges. Only DIFFERENCES within a component mean anything; the gauge is arbitrary.
Decomposition Lattice
gauge mean-zeroClient lens: dims drawn nodes, and clicking a covector bar re-expresses the potential against that asset. The server prune at 0.01 is what decided which nodes exist.
1 of 1 enumerated arcs earned ink: a node is drawn when it holds a live position or its residual clears 0.01. Everything else is implied and renders as scaffolding.
Feasible Shadow
About this projection
Basis pair = the two loudest assets in the fit, by total absolute flow (cre8/coin, cre8/item/chess/piece/15); 1 of 1 live positions project onto it. With divisible resting depth each position is a segment in bundle-flow space and the market's reachable set is their Minkowski sum (a zonotope), so this 2-D shadow is the EXACT zonogon, not a sketch: projections of polyhedra are polyhedra. It is a shadow, though, so it can be strictly wider than any single quote in that sub-market, and it models neither all-or-nothing fills, fees, settlement conservation, nor integral sizes.
Cleared Flow
window
60 cleared transactions scanned, but no net flow survives on the asset graph (, plus 118 at the window edge with only one side inside the 60-transaction scan). A bilateral swap contributes equal and opposite legs, so a round of pure two-party exchange nets to zero here; that is a real statement about the round, not a missing read.
Clearing Certificate
- matcher
- v4
- graph revision
- 1
- tier
- structural
- claims
- 3
- disclosures
- 19
Enumeration scope: no enumeration happened at all (no directed cycle: the containment graph has no cyclic strongly connected component, so the matcher short-circuited before enumerating any template). A certificate over such a round covers nothing: read it as 'nothing was swept', never as 'nothing was found'.
| Commitment | Value | What it binds |
|---|---|---|
| Content hash | 690016a63001bd600914fdfa0d60dfb23870333d33b0beb4a88bc5477a880ad2 | sha256 over the canonical preimage. This is the line to compare between two runs. |
| Claim Merkle root | 955b2955b4be49ea52bb9561fb2f380d7cdd1a6eac79f894a03995deb6be03cb | one leaf per claim section. |
| Tier reference | structural:955b2955b4be49ea52bb9561fb2f380d7cdd1a6eac79f894a03995deb6be03cb | what a receiver stores or gates a tier floor on. |
| Positions digest | 424081fe5dab5b64d639a7d6dddd765c24cd449ae3c603530cdef37fee686477 | the position set the claims were checked over. |
| Arc-map digest | f1833d62a26d6be1fdb178ccf4c719eaf6f3d150d03b54ed3f9d8113913d4760 | the arc map the routes were resolved against. |
| Routes digest | e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855 | the routes this round emitted. |
| Potential digest | aa0a534d9ef5cc71c5ea4311cbd70164653775d0648e76377149a06497b4e6a7 | the price covector the reduced costs were measured against; absent means no π-relative claim is bound. |
| Legs digest | — absent | the cleared legs the conservation claim was checked over; absent means none were supplied. |
Feasibilityout of scope: not checked, NOT a pass
STRONG: structural, and re-checkable from the round's own arcs and positions with no prices involved.
Structural feasibility of the routes this round emitted, recomputed from (arc_map, positions, routes) with no prices involved: every filed arc is a faithful projection of the position it names in the book, slot arcs resolve, each slot's offer covers its predecessor's want, the wrap closes the cycle, every listed position rides a qualifying arc, amounts are strictly positive, and no position is listed twice within a route or across routes. Covers ASSET-SET containment only — quantity sufficiency is a settlement-layer property and is out of scope here.
| Check | Verdict | Examined | Load-bearing | What was established |
|---|---|---|---|---|
| feasibility.position_id_unique | verified | 1 | yes | 1 position ids are pairwise distinct |
| feasibility.arc_backed_by_book | verified | 1 | yes | all 1 filed arc(s) carry the offer, want and peer their position posted |
| feasibility.arc_well_formed | verified | 1 | yes | all 1 filed arc(s) have indices derived from their own amount maps and are filed under their own arc key |
| feasibility.arc_filing_unique | verified | 1 | yes | each of 1 filed position(s) projects to exactly one arc |
| feasibility.route_shape | out of scope: not checked, NOT a pass | 0 | yes | no cleared route in this round |
| feasibility.route_id_unique | out of scope: not checked, NOT a pass | 0 | yes | no cleared route in this round |
| feasibility.arc_resolution | out of scope: not checked, NOT a pass | 0 | yes | no cleared route in this round, so no arc to resolve |
| feasibility.cycle_closure | out of scope: not checked, NOT a pass | 0 | yes | no cleared route in this round, so no cycle to close |
| feasibility.slot_membership | out of scope: not checked, NOT a pass | 0 | yes | no position was listed by any cleared route |
| feasibility.volume_positivity | out of scope: not checked, NOT a pass | 0 | yes | no position was listed by any cleared route |
| feasibility.intra_route_disjoint | out of scope: not checked, NOT a pass | 0 | yes | no cleared route in this round |
| feasibility.cross_route_disjoint | out of scope: not checked, NOT a pass | 0 | yes | no position was listed by any cleared route |
arc_records_filed 1arcs_in_round 1book_positions 1distinct_listed_positions 0listed_position_slots 0positions_reused_across_routes 0routes_examined 0slots_examined 0
Conservationout of scope: not checked, NOT a pass
STRONG: an integer ledger identity over the cleared legs, checked at zero tolerance. The decomposition identities riding alongside it are float algebra and are marked as not load-bearing.
Value conservation of the cleared legs, in i64 with zero tolerance and NO float tolerance on anything load-bearing: no leg amount is non-positive, no position delivers more of an asset than it posted in its own offer (totals in checked arithmetic — an i64 overflow is its own finding, never a saturated value that lands back on the cap), and no leg delivers an asset the receiving position did not want. Where several rows share one position id the TIGHTEST posted amount is the cap and an asset counts as wanted only if every such row wants it, so the verdict is a function of the digested book and not of the vector's order. This is a cap on what each position GIVES; it is NOT a check that anyone received the quantity they posted a want for (see LMC-D034), and not a price-relative value balance. The Hodge decomposition's algebraic identities ride along as non-load-bearing checks over whatever asset subset the decomposition priced. PER-ASSET DIVERGENCE IS RECORDED AS MEASURED AND NEVER ASSERTED TO BE ZERO — a non-zero divergence is the correct output of an open chain, not a violation.
| Check | Verdict | Examined | Load-bearing | What was established |
|---|---|---|---|---|
| conservation.leg_amount_positive | out of scope: not checked, NOT a pass | 0 | yes | no cleared legs were supplied to this call site; the matcher stage runs before routes_to_transactions, so conservation cannot be evaluated here |
| conservation.no_over_delivery | out of scope: not checked, NOT a pass | 0 | yes | no cleared legs were supplied to this call site; over-delivery against posted offers cannot be evaluated here |
| conservation.leg_authorised_by_route | out of scope: not checked, NOT a pass | 0 | yes | no cleared legs were supplied to this call site; no leg could be traced to the route that authorised it |
conservation_inputs_supplied nolegs_examined 0
No blocking templateout of scope: not checked, NOT a pass
WEAK: relative to the ENUMERATED template set only. This is not global optimality: the matcher enumerates and fills rather than solving an LP, so a template that was never enumerated was never checked, and no certificate tier can repair that: the gap is one of scope, not of proof strength.
Under the fitted price covector π, no template that was ENUMERATED and fully resolvable into a route, but not cleared, blocks this round. The skipped set splits by WHY it was skipped and the two halves carry different tests. A route's surplus is its negated holonomy −H: positive when the ring offers more than it wants and so can honour every limit, negative when no fill can. A skipped route sharing no position with anything cleared could have been added for free, so it must have surplus ≤ ε — no_blocking.reduced_cost. A skipped route that would have displaced cleared routes is judged by no_blocking.no_improving_swap: its surplus must not exceed the summed surplus of the routes it displaces, which is a 1-swap blocking-coalition test, NOT a claim that it was unprofitable. A skipped route whose settlement could not pay every member its full want is outside both domains: clearing it would settle somebody below their posted price. Both tests are Farkas-flavoured and both are explicitly relative to the enumerated template set: a template never enumerated is never checked, and one that failed structurally has no arc sequence and therefore no reduced cost at all. THIS IS NOT GLOBAL OPTIMALITY — the swap test is local (single swap), the enumeration is bounded, and the matcher is enumerate-and-select, not an LP solved to optimality. No strengthening of this certificate can change that.
| Check | Verdict | Examined | Load-bearing | What was established |
|---|---|---|---|---|
| no_blocking.price_fit_stationary | verified | 1 | yes | π is stationary for the residual it reports over 1 position(s): the largest first-order log-price move is 2.220446049250313e-16 ≤ τ 1e-6 (0 unpriceable position(s) excluded; 0 asset(s) priced at the boundary π → 0, where every residual they touch is settled to within 1e-8; the producer reported converged=true) |
| no_blocking.skipped_structural | out of scope: not checked, NOT a pass | 0 | no, does not gate the claim | every enumerated template assembled into a route |
| no_blocking.reduced_cost | out of scope: not checked, NOT a pass | 0 | yes | VACUOUS: the conflict-free skipped domain is EMPTY. 0 template(s) were enumerated, 0 route(s) cleared, and of the 0 assembled-but-skipped route(s) every one shares a position with something that cleared, so none could have been added for free. Those are adjudicated by no_blocking.no_improving_swap instead; this check had nothing to evaluate and establishes NOTHING about this round. |
| no_blocking.no_improving_swap | out of scope: not checked, NOT a pass | 0 | no, does not gate the claim | VACUOUS: no assembled-but-skipped route shares a position with anything that cleared, so no route was displaced and there is no swap to adjudicate. 0 route(s) cleared. This check establishes NOTHING about this round. |
cleared_routes_priced 0enumerator_reported_count 0price_components 1price_fit_converged yesprice_fit_stationarity 0priced_arcs 1priced_assets 2pricing_inputs_supplied yesreduced_cost_epsilon 0routes_cleared 0routes_matched 0skipped_conflict_free 0skipped_displacing 0skipped_limit_infeasible 0skipped_priceable_domain 0skipped_priced 0skipped_structural 0skipped_unpriced_arcs 0stationarity_boundary_assets 0stationarity_positions 1stationarity_tolerance 0stationarity_unpriceable 0templates_enumerated 0templates_skipped 0
This certificate does NOT cover (19 disclosure(s)):
- LMC-D001 Route ids are NOT content commitments. A route id is sha256 over the template's want sequence only, so two entirely different books that contain the same want cycle produce the same route id. Bind to this certificate's round_digest, never to a route id.
- LMC-D002 arc_sequence is ONE representative arc per slot, chosen by maximum offer volume with a reverse lexicographic tiebreak. position_sets is the strictly larger set that actually settles. Every derived field (index_sequence, face_ranks, offer_ranks, min_face_rank, lattice_span, flow_tier) describes the representative, not the settling set.
- LMC-D003 volume_per_slot and min_slot_volume are sums of raw offer amounts across heterogeneous assets and across every qualifying position. They are ranking heuristics, not values, not notionals, and not bounds on cleared value.
- LMC-D004 Want coverage is checked at the ASSET-SET level, not the quantity level. Settlement fills each slot greedily at min(supply, demand) per asset with no cross-slot balancing, so a route can be feasible here and still leave demand unmet downstream.
- LMC-D005 No self-dealing check. A single peer may occupy several slots of one route; nothing in the matcher or in this certificate requires distinct peer ids across slots.
- LMC-D006 The matcher reserves nothing. A qualifying arc is listed in full by every route that resolves it, and routes_to_transactions allocates each route independently over the same position list. Cross-route reuse is prevented only downstream, at the settlement chokepoint, and only for CONFIRMED routes — a different layer than the one this certificate covers.
- LMC-D007 Slot fill is volume-maximal, not price-optimal: the canonical arc is the deepest qualifying arc, never the best-priced one. Whether a different candidate arc at some slot would lower a cleared route's reduced cost is UNCHECKED.
- LMC-D008 Templates with two or more empty slots are discarded by match_templates without being recorded anywhere in the match result — they are not even reported as near-misses. They ARE inside this certificate's skipped census (templates_skipped / skipped_structural), because that census is a set-difference against the enumerated list rather than a reading of the matcher's own output. What is NOT available for them is WHICH slots were empty, or how close they came.
- LMC-D009 The price fit truncates each position side to the first 4 assets in LEXICOGRAPHIC order, not by volume. The 5th and later assets of a wide basket contribute no pricing evidence and the basket itself is mis-scaled in the refinement stage.
- LMC-D010 Log-prices are gauge-fixed to mean zero PER CONNECTED COMPONENT. Only ratios within one component mean anything; any absolute value figure is in that component's arbitrary gauge units, and comparing value across components is meaningless.
- LMC-D011 π is settled only to about 1e-6 in log-price in well-constrained directions and materially looser in weakly constrained ones. No π-relative epsilon below 1e-6 is defensible, and this module clamps to that floor.
- LMC-D012 arc_residuals and bundle_premiums are weighted-least-squares FIT RESIDUALS. They measure fit quality on an inconsistent book and are never conservation violations.
- LMC-D013 Per-asset divergence from the flow decomposition is reported AS MEASURED. A non-zero divergence is the correct output of an open chain, not a defect; an asset no priced leg touched is unmeasured, not a balanced zero.
- LMC-D014 This certificate is tier 'structural': a recomputable commitment. Anyone holding the round can recompute the canonical bytes, the content hash, and the claim Merkle root. That is NOT a proof, NOT a signature, and NOT peer attestation.
- LMC-D015 NO tier of this ladder can repair the no-blocking-template claim. Higher tiers raise confidence that the arithmetic ran and was not altered; that claim's weakness is a SCOPE limit relative to the enumerated set, which only a different matching algorithm (LP duality over the full template space) can close.
- LMC-D016 Arc keys are the string 'offer_index==want_index' and each index is ':'-joined asset names, so ':' and '==' are structural separators in an ASSET NAME position. An asset actually named with one of them is mis-parsed, and it is mis-parsed IDENTICALLY by the matcher and by this module's deliberately independent containment predicate — so the second implementation is no defence against it. The round digests are separator-escaped and are not affected; the feasibility predicates are.
- LMC-D020 NO ENUMERATION OCCURRED in this round (no directed cycle: the containment graph has no cyclic strongly connected component, so the matcher short-circuited before enumerating any template). This certificate covers nothing on the no-blocking-template axis: 'no blocking template was found' would be a false reading of 'no template was looked at'.
- LMC-D030 No cleared legs were supplied at this call site, so the conservation claim is OUT OF SCOPE. It is not a pass. Re-issue after routes_to_transactions to obtain a conservation verdict.
- LMC-D044 The enumerated-template list this claim is measured against is supplied by the caller (0 template(s)) alongside the enumerator's own reported count (0). Nothing here cross-checks the two, and no check is refuted when they disagree — a truncated list silently shrinks the skipped census and makes the weak claim look cleaner than the round was. Compare the templates_enumerated and enumerator_reported_count evidence counters yourself.
Self-consistency: the stored content hash and claim root recompute from the stored content, and every claim's verdict follows from its own checks. That is an integrity check on the ARTIFACT: it says the object was not edited after it was written. It is not a verification of the round: only re-running the checkers against the round's own positions, arcs, routes and legs (verify_clearing_certificate) does that, and this page deliberately does not, because the book it can read now is not the book that cleared.
Three verdicts, never two: a filled green pill is a predicate that ran over a non-empty domain and held; a filled red pill is a counterexample, named above; a hollow dashed pill is out of scope (the predicate did not run, or ran over an empty domain), and it is drawn hollow for the same reason an unmeasured asset on the flow map is drawn hollow. It is not a pass.
Liquidity Field
No gap shapes persisted yet. The matcher records a near-miss only when a route template comes up exactly one arc short, so an empty field means either no match run has happened against this store or every ring it walked was already complete. It does NOT mean the book is saturated. Run a match (LatticeMatch.Routes) over a book with an unfilled ring and the wanted ads appear here.