I can give an implementable definition, but I could not verify the requested 14-day sale ledger or block-pinned prices. The accessible pages provide incomplete marketplace snapshots; RPC, explorer and API attempts did not yield the underlying data. Reporting numerical “fair floor” or wash-trade findings from those snapshots would be unsupported.

Below is the proposed method and the precise status of the identity.md application.

1. Proposed fair-floor definition

I would define a fair floor as a conservative estimate supported by independent executable asks and recent arm’s-length sales, with a separately verified liquidation bid as a lower bound. It is an estimator, not a guarantee that an NFT can be bought or sold at that price.

The following thresholds are my proposed policy parameters, not established industry standards.

Snapshot and price conventions

Choose one finalized Ethereum block (B), record its hash and timestamp (T), and use:

[ W=(T-14\times86400,\ T]. ]

Read all on-chain state at (B). Archive signed off-chain orders, their retrieval timestamps and the marketplace coverage list: a block number alone cannot reconstruct an off-chain order book.

Use ETH-equivalent amounts, treating ETH and canonical Ethereum WETH as equal units. Record the actual payment currency too. Define:

• Ask: total buyer payment, including mandatory marketplace fees and royalties, excluding gas.
• Sale: total consideration attributable to that NFT, including mandatory fees, excluding gas.
• Bid: seller’s net proceeds after mandatory deductions, excluding gas.

Keep sales paid in other currencies in the ledger, but exclude them from this estimator unless a separate conversion policy is specified. Never divide a bundle price equally among NFTs without evidence of that allocation.

Validate listings before using them

An eligible ask must satisfy all of these at (B):

1. The seller owns the NFT, and the necessary approvals exist.
2. Signature, counter/nonce, cancellation status, remaining quantity, time bounds and any zone restrictions permit fulfillment.
3. An ordinary unrelated buyer can fulfill it; exclude private and conditional trades unavailable to that buyer.
4. Simulating fulfillment against block (B) succeeds with the advertised total payment.
5. The signed order was retrieved within five minutes of the snapshot.

An old order can still be valid; age alone is not grounds for exclusion. Revalidation addresses staleness. OpenSea’s documentation specifically distinguishes events from order invalidation/revalidation and directs users to order-status checks. OpenSea events reference

Deduplicate by NFT, retaining its cheapest eligible ask. Define the naive floor:

[ N_B=\min_i a_i. ]

For the robust ask estimate, group sellers into ownership entities using documented common control, retaining one cheapest NFT per entity. Mere funding from the same exchange does not establish common control.

Sort the resulting prices (a_{(1)}\le\cdots\le a_{(m)}). Require at least five entities:

[ A_B=\operatorname{median}(a_{(1)},\ldots,a_{(5)})=a_{(3)}. ]

This is an equally weighted, seller-diversified lowest-five estimator. Compared with a lowest-five mean, its value is less sensitive to the magnitude of one extreme price. High troll listings generally fall outside the selected five. Low executable listings remain visible in (N_B), even when they do not set (A_B).

Construct and wash-filter the sale ledger

Enumerate collection transfers, then join transaction receipts, marketplace settlement events and payment traces. A transfer is not proof of a sale: ERC-721’s Transfer event covers ownership changes “by any mechanism,” including minting and burning. ERC-721 specification

Record one row per NFT sale, identified by transaction hash, settlement event/order identifier and token ID. Record the settlement protocol separately from any known frontend: an aggregator displaying a transaction does not establish which marketplace executed it.

Assign these reproducible flags:

Flag	Proposed test
SAME_CONTROL	Buyer and seller are the same address or belong to a documented common-control entity.
CIRCULAR	Within a 30-day lookback ending at (B), the NFT follows a time-ordered ownership path returning to a prior entity. Flag all sale edges in that closed path.
SELLER_FUNDED	In the 72 hours before payment, seller-origin ETH/WETH reaches the buyer through at most three intermediary addresses and funds at least 50% of the consideration.
UNKNOWN	Required ownership history, settlement allocation or funding traces are missing.
NO_FLAG_FOUND	Required checks completed without triggering the above tests. This does not mean “proven organic.”

For the funding test, process ETH/WETH transfers and wraps in transaction order, using FIFO attribution of incoming funds to outgoing funds. Seed attribution at the seller entity; avoid counting the same funds twice. Stop attribution at documented pooled exchanges, bridges and custodians, where individual provenance cannot be established. Publish the registry and traced paths with the result.

Exclude the first four categories from the estimation sample, but retain every row in the published ledger. Circular trading and seller financing are suspicion indicators; they can also occur in legitimate transactions.

To limit one trader’s influence, sort remaining sales newest first, breaking ties by transaction index and log index. Greedily retain a sale only if:

• its token has not already been retained; and
• neither participant entity has already appeared in a retained sale on that UTC date.

For retained sale (j), assign:

[ w_j=2^{-(T-t_j)/(3\times86400)}. ]

Require at least ten retained sales, five distinct buyer entities and five distinct seller entities. Let (S_B) be their weighted median: the smallest price whose cumulative weight reaches half the total.

The median limits price-outlier influence; the three-day half-life reduces the influence of old sales. Neither defeats coordinated traders controlling enough apparently independent wallets.

Bids: distinguish item-specific demand from a collection lower bound

For each transferable token (i), calculate:

[ b_i=\max{\text{net proceeds from each eligible executable offer for }i}. ]

Include item offers, applicable trait/collection offers and on-chain buying contracts. Validate funding, allowance, expiration, remaining quantity and actual fulfillment at (B). OpenSea notes that collection offers can be restricted by traits, so their applicability must be checked. OpenSea collection-offer explanation

Report both:

[ B_{\mathrm{top}}=\max_i b_i,\qquad L_B=\min_{i\in U_B}b_i, ]

where (U_B) is the published set of existing transferable collection tokens. With complete offer coverage, assign zero to a token with no executable bid; with incomplete coverage, mark the result unknown.

• (B_{\mathrm{top}}) is the best item-or-collection bid anywhere.
• (L_B) is a lower bound applicable to any one token in the defined universe.
• Neither implies enough liquidity to sell the whole collection. Shared bidder funds must not be counted repeatedly as depth.

For IMDSTR at 0x0000198C940D8cD70Cb9ACeC5E3af8216ac57d2F, include only proceeds achievable through an executable purchase path. Its treasury balance, previous purchase price or relisting multiple is insufficient. Verify spending limits, permissions, token eligibility, available funds and execution conditions. A purchase requiring an uncommitted keeper action is conditional demand, not a firm bid.

Chosen formula

When both (A_B) and (S_B) meet their minimum samples, bid coverage is sufficient to establish (L_B), and (L_B\le A_B):

[ \boxed{F_B=\max\left(L_B,\ \min(A_B,S_B)\right)}. ]

The minimum requires both current supply and recent trading to support a higher valuation. The bid provides an executable lower bound within its stated scope. This deliberately produces a conservative estimate and can lag a rapidly rising market.

If inputs are missing, report fair floor unavailable, alongside whatever components are established. If (L_B>A_B), report a crossed or inconsistent snapshot and investigate before publishing (F_B).

Publish:

[ \text{robust spread}=A_B-L_B,\qquad \text{relative spread}=\frac{A_B-L_B}{(A_B+L_B)/2}. ]

Also publish (N_B-L_B), the naive ask versus the common liquidation bound. Do not subtract a rare-token-only top bid from an unrelated floor listing and call it a tradable spread.

2. Application to identity.md

Target contract: 0x0000eC93127BAA929E58E97dd0095A2BFb38ec1D.

For a reproducible date-only interpretation, the intended reporting interval is 2026-09-11 00:00:00 UTC through 2026-09-25 00:00:00 UTC, start inclusive and end exclusive. Resolving those boundaries to Ethereum blocks was unsuccessful.

The accessible OpenSea activity page displayed sales and USD amounts but did not expose the transaction hashes, sale timestamps, settlement blocks or ETH consideration needed to establish the requested ledger. Its retrieved page also repeated activity rows, making naive page-row counting unsafe. Identity MD activity

The OKX page identifies the requested collection contract, but its displayed last-sale snippets are not a timestamped, complete transaction history. Identity MD on OKX

Accordingly, the defensible results are:

Requested output	Result	Block pin
Every sale in the 14-day interval	Not established; complete ledger unavailable	Unavailable
Per-sale wash flags	Not assessed; funding and ownership histories unavailable	Unavailable
Naive executable floor (N_B)	Unavailable	Unavailable
Robust ask (A_B)	Unavailable	Unavailable
Wash-filtered sales median (S_B)	Unavailable	Unavailable
Fair floor (F_B)	Unavailable	Unavailable
Best bid (B_{\mathrm{top}}) and common bound (L_B)	Unavailable, including IMDSTR contribution	Unavailable
Bid-ask spread	Unavailable	Unavailable

This is an incomplete empirical application, not evidence that there were zero sales or no bids. I cannot supply one verified row per sale or attach genuine block numbers to these measures from the data retrieved.

3. Access limitations and uncertainty

During this research:

• Direct requests to two Ethereum RPC endpoints failed with DNS-resolution errors in the execution environment.
• The collection’s Etherscan token page returned HTTP 403; the strategy contract and Blockscout attempts did not return usable contract/state data.
• The OpenSea collection-events API request was inaccessible through the browsing tool. Its official documentation describes timestamp filtering and cursor pagination—the data path needed for a complete marketplace event history—but I did not retrieve those records. OpenSea analytics and events documentation
• I obtained no complete signed order book, block-pinned fulfillment simulations, funding traces or verified IMDSTR configuration.

Even with those inputs, hidden common ownership, off-chain financing and pooled-fund transfers remain uncertainty sources. Recent circular trades can also become detectable only after the reporting block. A completed report should therefore publish the raw sale ledger, exclusion reasons, entity evidence, order coverage and block hash alongside the estimated price.

---

Sources offered but not confirmed by this machine:

• https://web3.okx.com/nft/collection/eth/identitymd-1 — does not contain the quoted text

I can specify an implementable fair-floor estimator, but I could not verify the requested 14-day sales ledger or block-pinned prices. The accessible marketplace snapshot shows a 1.99 ETH displayed floor; it is not enough to establish an executable floor, fair floor, or best bid. I have not invented sales, wash-trade flags, or block numbers. OpenSea collection

The following thresholds are my proposed methodology, not a published standard or a calculation completed for this collection.

1. Definition and calculation

Define “fair floor” as a conservative estimate of an ordinary collection member’s value, supported by independent executable supply, screened transactions, and immediately available demand. It is an estimator, not a guaranteed liquidation price.

Fix a finalized Ethereum block (B), record its hash and timestamp (T), and use sales with timestamps in ((T-14\text{ days},T]). Archive the signed orders used in the calculation, their retrieval timestamps, and the validation results at (B). A block number alone cannot document which off-chain orders were available.

Use ETH-equivalent amounts: ETH plus canonical WETH at 1:1 as an explicit accounting convention. Exclude gas; include compulsory marketplace fees and royalties in acquisition prices. Record seller-net proceeds separately.

Validate and consolidate listings.

For each listing, require:

• Correct collection and token, current ownership and approvals.
• Valid signature, nonce/counter, start/end time, remaining quantity, and cancellation status.
• Successful simulated fulfillment against state at (B), with no private-buyer restriction or unavailable prerequisite.
• Retrieval within five minutes of the snapshot; reject unvalidated cached entries.
• One cheapest executable order per token across venues.

An old listing that still passes these checks remains eligible. Age alone does not make it stale. ERC-721 specifies ownership and approval interfaces; marketplace orders additionally encode what is offered and what must be provided in return. ERC-721 specification, Seaport documentation

Let the naive executable floor be:

[ N_B=\min_{\ell\in L_B} a_\ell ]

where (a_\ell) is the full acquisition cost.

For the robust ask component, cluster sellers by documented common control, retaining the evidence and clustering rules. Do not equate a shared exchange withdrawal source or shared marketplace operator with common ownership. Take each cluster’s cheapest listing, sort these prices, and select the lowest five:

[ a_{(1)}\leq\cdots\leq a_{(5)},\qquad A_B=a_{(3)}. ]

Require five seller clusters; otherwise (A_B) is unavailable. This is the equal-depth weighted median of the lowest five independent sellers. Two extreme entries within those five cannot alone determine it. A lowest-five arithmetic mean is more sensitive to extreme prices; an unrestricted quantity-weighted measure lets one seller supply most of the apparent depth.

Build and screen the sales ledger.

A transfer is not automatically a sale: ERC-721’s Transfer event covers ownership changes by any mechanism. Decode marketplace settlements, payment transfers, refunds and, where necessary, execution traces. Deduplicate by transaction hash, settlement event and token. Report the settlement protocol separately from the frontend; a Seaport settlement does not, by itself, establish which frontend originated it. ERC-721 specification, Seaport documentation

Each sale row should contain:

block | tx hash | timestamp | token | buyer | seller | gross ETH | seller-net ETH | protocol/frontend | wash flag | evidence

For bundles lacking explicit per-token allocations, retain a row per token with price unavailable—bundle allocation unknown. Do not invent equal allocations.

Use these deterministic screening flags:

Flag	Proposed rule
Same controller	Buyer and seller are the same address or belong to the same documented control cluster.
Circular ownership	The same token returns to a prior owner cluster within seven days through a cycle containing 2–4 sale edges. Flag every sale in that cycle.
Seller-funded buyer	During the 72 hours preceding settlement, seller-origin ETH/WETH reaches the buyer, directly or through at most two intermediary wallets, amounting to at least 80% of the acquisition cost.
Payment round trip	At least 80% of acquisition cost returns from seller to buyer within 72 hours after settlement.
Unknown	Required payment, control or funding history is unavailable.

For the funding tests, trace funds chronologically using FIFO attribution, combine ETH/WETH wrapping as the same value, and stop attribution at identified exchanges, bridges and pooled protocols. Freeze and publish the service-address list used. These are suspicion rules, not proof of intent; legitimate resale or financing may trigger them. Address pseudonymity also permits undetected common control, a limitation emphasized in NFT wash-trading research. von Wachter et al.

Fetch at least seven days before the sales window to detect boundary-crossing cycles. Never use observations after (B); recent rows have incomplete forward-looking cycle/refund checks and must be marked provisional.

Exclude flagged and unknown rows from the estimator, while keeping them in the ledger. Among eligible sales, retain only the latest sale per token, then only the latest sale per unordered buyer/seller-cluster pair. Break timestamp ties by block, transaction index and log index.

For each remaining sale (s), set:

[ u_s=2^{-(T-t_s)/(3\text{ days})}, ]

[ w_s=\frac{u_s} {\max\left( 1,\sum_{r:,\mathrm{buyer}(r)=\mathrm{buyer}(s)}u_r, \sum_{r:,\mathrm{seller}(r)=\mathrm{seller}(s)}u_r \right)}. ]

This caps each buyer’s and seller’s total weight at one. Let (M_B) be the weighted median gross price: the smallest price at which cumulative sorted weight reaches half the total.

Require at least ten eligible sales, five buyer clusters and five seller clusters. Otherwise (M_B) is unavailable. This component reflects completed trades but can lag a changing market and can still overrepresent premium tokens.

Measure executable bids, including the strategy.

For every live token (i), let (b_i) be the highest seller-net proceeds from an executable item offer, qualifying collection offer, or permissionlessly callable on-chain purchase mechanism at (B). Require valid orders, sufficient funds and allowances, satisfied criteria, and successful settlement simulation. Do not count self-bids.

Report both:

[ B_{\mathrm{item}}=\max_i b_i,\qquad B_{\mathrm{floor}}=\min_i b_i. ]

The highest item offer is not a lower bound for the whole collection. (B_{\mathrm{floor}}) describes one-token liquidation coverage across the collection; it does not imply that all tokens could simultaneously sell at that price. Report funded quantity and shared-capital constraints separately. Missing bid coverage makes the collection-wide bound unavailable, not zero.

For the supplied IMDSTR address, inspect verified implementation code and state, purchase permissions, eligible tokens, pricing limits, available funds, cooldowns and competing commitments. Simulate its purchase path. A contract balance, historical purchase, intended future sweep, or relisting multiple is not itself an executable bid. A real strategy purchase is not automatically a wash trade; its subsequent listing counts as one seller cluster.

My chosen formula is:

[ \boxed{F_B=\max!\left(B_{\mathrm{floor}},,\min(A_B,M_B)\right)} ]

Require all three components and (B_{\mathrm{floor}}\leq A_B); otherwise publish insufficient data or crossed/inconsistent snapshot, respectively. The minimum prevents unsupported asks alone from raising the estimate; the maximum respects verified collection-wide liquidation demand. Report the three inputs alongside the result so the policy choice remains visible.

Report spreads as:

[ \Delta_{\mathrm{naive}}=N_B-B_{\mathrm{floor}},\qquad \Delta_{\mathrm{fair}}=F_B-B_{\mathrm{floor}}. ]

These are before gas and use acquisition-cost asks versus seller-net bids. No estimator here is Sybil-proof: undiscovered coordinated wallets can corrupt both samples.

2. Application to identity.md

Target: Ethereum 0x0000eC93127BAA929E58E97dd0095A2BFb38ec1D.

For a reproducible calendar cutoff, I would use 2026-09-11 00:00 UTC through 2026-09-25 00:00 UTC, selecting the last finalized block before the latter timestamp and calculating the precise trailing 14 days from that block. I could not retrieve that block.

The accessible OpenSea page displayed a 1.99 ETH floor, a $5,349.98 floor valuation, a $4,921.71 top offer and 98 listed items. It also showed individual listing cards, but the retrieved content did not establish signed-order validity or a block snapshot. OpenSea collection

Requested result	Verified result	Read block
Naive executable floor (N_B)	Unavailable; displayed floor only: 1.99 ETH	Unavailable
Fair floor (F_B)	Unavailable	Unavailable
Best executable item bid (B_{\mathrm{item}})	Unavailable	Unavailable
Collection-wide bid bound (B_{\mathrm{floor}})	Unavailable	Unavailable
Executable bid–ask spread	Unavailable	Unavailable

The displayed dollar figures differ by $428.27, but that arithmetic is not a verified executable spread. It does not establish offer eligibility, seller-net proceeds, funding, or synchronized observations. Displayed inputs

Sale-by-sale ledger: not established. I cannot provide one row per sale, sale prices, marketplace attribution or defensible wash flags from the data obtained. This means unknown sales coverage, not “zero sales” and not “no wash trading.”

3. Access limitations and uncertainty

My RPC requests to PublicNode and LlamaRPC failed with DNS-resolution errors in this environment. The attempted Blockscout transfer endpoint and Etherscan strategy-contract page were also inaccessible through the available retrieval routes. I consequently obtained neither the full transfer/settlement history nor funding traces, strategy code/state, or a finalized snapshot block.

The accessible OKX page explicitly associates its collection with the requested contract, but its retrieved FAQ says there are no active listings, while the OpenSea snapshot displays listed items. Those pages do not establish synchronized observations and cannot resolve the executable market. OKX collection, OpenSea collection

Completing the numerical application requires an archive-capable RPC with receipts and traces, complete paginated settlement history, and retrievable signed listings/offers validated at a common block. The methodology above is complete; the requested empirical audit remains incomplete.

---

Sources offered but not confirmed by this machine:

• https://docs.opensea.io/docs/seaport — did not return a page
• https://web3.okx.com/nft/collection/eth/identitymd-1 — does not contain the quoted text

Conclusion

A defensible “fair floor” is not the cheapest listing. It is a robust estimate derived from executable listing depth and wash-filtered sales, bounded below by the highest immediately executable bid.

I could define that measure precisely, but I could not truthfully complete the requested block-pinned IdentityMD calculation from the data interfaces available in this environment. In particular, I could not retrieve a complete order book or decode every marketplace fill for the 14-day window. I therefore do not invent sale rows, block numbers, bids, or a fair-floor value.

1. Implementable definition

Evaluate everything at Ethereum block (B).

Eligible asks

For every marketplace order (i), obtain its all-in buyer price (a_i), including mandatory fees. Retain it only if, at (B):

1. the order is signed and unexpired;
2. its maker owns the token;
3. the marketplace conduit has approval;
4. it is fillable for ETH or WETH;
5. the token is not simultaneously committed to an earlier competing order;
6. the order was created or revalidated within seven days.

Sort eligible asks by price. Keep at most one token per beneficial owner in the initial depth calculation so one wallet cannot fabricate depth.

Find the first cluster of at least three distinct sellers whose prices all lie within 20% of the cluster median. Lower isolated asks are labelled troll/outlier listings and skipped.

Let (A_5) be the weighted median price of the cheapest five tokens in that cluster, or all tokens if fewer than five survive. Each distinct seller receives total weight one, divided among that seller’s included tokens.

This is preferable to the literal minimum: a marketplace snapshot recently reported 0.68 ETH, while a later Blur crawl showed 1.78 ETH—large enough variation to make a single displayed minimum unsafe as an oracle. CoinStats Blur

Eligible sales

Decode marketplace settlement events over blocks whose timestamps are in ([t_B-14\text{ days},t_B]). Count one row per ERC-721 exchanged for ETH/WETH; allocate bundle consideration pro rata only if the protocol supplies explicit item amounts, otherwise exclude the bundle.

For each sale, construct seller and buyer funding/ownership clusters. Flag and exclude a sale when any condition holds:

• seller and buyer are identical or controlled by the same address;
• the token returns to a previous owner within seven days;
• buyer funds originated from the seller, a seller-controlled address, or the sale proceeds in the preceding 72 hours;
• a directed cycle of length 2–4 exists between the participating clusters within 30 days;
• repeated reciprocal trading occurs between the same clusters;
• consideration is returned to the buyer outside normal marketplace refunds;
• the transaction is a private transfer mislabeled as a sale.

Do not automatically reject a purchase by IMDSTR merely because it is systematic; treat the strategy as a disclosed market participant. Exclude it only if the funding graph or circular ownership rule actually triggers.

Give surviving sale (j), price (p_j), age (d_j) days, weight

[ w_j=2^{-d_j/7}/n_{\text{buyer-cluster},j}, ]

where (n_{\text{buyer-cluster},j}) is that buyer cluster’s number of eligible purchases in the window. Let (S_{14}) be the weighted median.

A previous on-chain IdentityMD calculation found a 1.09 ETH lowest Seaport fill and another found a 1.29 ETH average in their respective pinned 24-hour windows, illustrating why the calculation must state its block interval rather than reuse a “recent” figure. IMD agent result IMD average-sale result

Executable bid

For every item offer, collection offer, and contract buyer, calculate net ETH/WETH payable to a seller of an ordinary eligible token. At (B), verify:

• signature and expiry;
• allowance and balance, or escrowed balance;
• criteria proof;
• remaining quantity;
• token eligibility;
• fees and royalties affecting seller proceeds.

For IMDSTR at 0x0000198C940D8cD70Cb9ACeC5E3af8216ac57d2F, use a successful eth_call at block (B) for its actual purchase method and an ordinary IdentityMD token. A contract balance alone is not a bid.

Let (Q_B) be the maximum verified net proceeds from these routes.

Fair floor

When both components have adequate observations—at least three distinct eligible listing sellers and five non-wash sales from at least three buyer clusters—define

[ F_B=\max\left(Q_B,; \min\left(A_5,\exp(0.60\ln A_5+0.40\ln S_{14})\right)\right). ]

Thus an executable bid is a hard lower bound, thin cheap asks cannot dominate, and old sales have declining influence.

Fallbacks:

• Insufficient sales: (F_B=\max(Q_B,A_5)), labelled “ask-based, low confidence.”
• Insufficient asks: (F_B=\max(Q_B,S_{14})), labelled “sales-based, low confidence.”
• Neither: report only (Q_B); do not call it a floor.
• Spread: (A_5-Q_B) ETH and ((A_5/Q_B-1)\times100%), provided (Q_B>0).

2. IdentityMD application and audit result

Proposed read block: 26,052,604, timestamp 2026-09-25 06:13:59 UTC. The explorer page identifies that height and timestamp; Ethereum’s JSON-RPC specification explains that historical state calls must supply the desired block parameter. Ethscan block feed Ethereum JSON-RPC specification

Requested 14-day interval: 2026-09-11 06:13:59 UTC through block 26,052,604.

Required result at block 26,052,604	Result
Complete sale table	Unavailable—not safely reconstructable from accessible data
Naive active-listing floor	Not verified at the pinned block
(A_5)	Not computable
(S_{14})	Not computable
Best executable bid (Q_B)	Not computable
Fair floor (F_B)	Not computable
Bid–ask spread	Not computable

This is not a claim that no trades or orders existed. Routescan documents a keyless ERC-721-transfer endpoint returning block number, transaction hash, addresses, timestamp, token address, and token ID, but that endpoint was inaccessible from this session; moreover, transfers alone do not identify consideration or marketplace. Routescan ERC-721 API

The partial marketplace material is insufficient:

• OKX exposed last-sale values for several tokens—such as #1565 at 0.6489 ETH and #855 at 0.453 WETH—but did not expose the block, transaction, counterparties, or a complete 14-day result set in the accessible page. Those entries therefore cannot serve as the requested sale ledger. OKX collection page
• OpenSea’s indexed page exposed a displayed collection floor but not a reproducible block-pinned order book. OpenSea activity
• A third-party collection page describes 2,000 items, whereas CoinStats reported 5,000, another reason to avoid trusting aggregator metadata without direct contract verification. OKX collection page CoinStats

3. Uncertainty and missing access

The missing inputs were:

• complete Seaport, Blur, LooksRare, and other settlement logs and transaction receipts for the interval;
• historical order signatures and active/cancelled status at block 26,052,604;
• maker ownership and approval checks at that block;
• collection and item bids with balances and allowances;
• IMDSTR’s verified ABI/source and a historical simulation of its buy function;
• counterparty funding traces needed for seller-funded-buyer and circular-sale detection.

Consequently, any numeric “fair floor,” best bid, spread, or supposedly complete per-sale table would be fabricated. The correct reproducible next step is to query an archive Ethereum node at block 26,052,604, decode every qualifying settlement, then publish the transaction hashes and the exact order set alongside the calculation.

---

Sources offered but not confirmed by this machine:

• https://routescan.io/docs/api/tokens — did not return a page
• https://coinstats.app/nft/0x0000ec93127baa929e58e97dd0095a2bfb38ec1d/ — does not contain the quoted text
• https://web3.okx.com/zh-hans/nft/collection/eth/identitymd-1 — does not contain the quoted text
• https://explorer.imd.fun/agents/85 — does not contain the quoted text
• https://explorer.imd.fun/agents/912 — does not contain the quoted text
• https://ethscan.org/ — does not contain the quoted text

identity.md fair floor, read at block ≈26,055,887 (2026-09-25 17:14 UTC)

Short answer: I estimate the fair floor at about 1.99 ETH, with a likely range of 1.95–2.00 ETH. The naive floor is 1.99 ETH, the best bid I could see on-chain is 1.78 WETH, and the spread is about 0.21 ETH (roughly 11%).

What's missing: I did not get the full 14-day sale list. The only sales data I could reach was a public block explorer (Blockscout), and each price had to be looked up one transaction at a time. The sale table below covers only the last roughly 5 hours (blocks 26,054,425–26,055,887). At about 45 sales a day (inferred from OpenSea's 24h volume of about $307k), the full window likely holds 300–600 sales, so most of them are not listed here.

1. Method: fair floor (FF)

The inputs, all read at one block b:

• Fillable listings, L. Active listings for the whole collection (not trait listings), priced in ETH or WETH, that meet all of these:

  • the lister still owns the token at block b;
  • the marketplace is still approved to move it (isApprovedForAll or getApproved);
  • the listing expires more than 1 hour after b;
  • it was created or refreshed within the last 30 days (older ones count as stale);
  • the token has not been flagged on the marketplace.

  Keep at most 2 listings per owner, counting known linked wallets (same funder, same ENS) as one owner.

• A₁, A₃ and A_k. Sort L by price p₁ ≤ p₂ ≤ …. A₁ is the lowest price and A₃ the third lowest. A_k is the average of the lowest k = 5, which is the cost per item of buying 5. This depth weighting stops one fake or throwaway listing from moving the floor.

• Clean sales, S. Sales in the last 7 days. If there are fewer than 8, extend to 14 days. First remove wash trades (defined below). Then remove likely rarity-premium sales priced above 1.5 × A_k.

• M_s. A time-weighted median of S, with each sale's weight wᵢ = 2^(−age_hours/72) (a 72-hour half-life).

• B, the best bid you could actually sell into: the largest of:

  • the top collection offer, capped at what the bidder can pay: min(offer, the bidder's WETH balance, the bidder's WETH approval to the marketplace);
  • the top item offer, only if it applies to any token;
  • on-chain buyers such as the IMDSTR strategy, capped at min(getMaxPriceForBuy(), the contract's ETH balance).

  Each is taken after marketplace and creator fees.

Formula:

raw  = 0.5·A_5 + 0.5·M_s
FF_b = min( A_3 , max( B , raw ) )
spread_b = A_1 − B ;  spread% = (A_1 − B) / ((A_1 + B)/2)

Why these bounds:

• Upper bound A₃: you can never pay more than it costs to buy three real listings.
• Lower bound B: you can always sell into the best bid that can actually be paid.
• The sales term reduces the effect of listing games, such as delisting to push the floor up or fake cheap listings that can't be filled.
• I rejected using the lowest listing A₁ alone (easy to fake with one listing), sales alone (slow to react, can be washed) and bids alone (too low, and the IMDSTR bid is mechanical).

Wash-trade rule. Drop a sale if any of these hold:

• (a) the buyer and seller are the same, or are linked through the funding checks in (c) and (d);
• (b) the token returns to a previous owner within 30 days (a circular trade);
• (c) the buyer received more than 50% of the purchase price from the seller, or from a common funder, within 7 days before the sale;
• (d) either side was funded through a cross-chain relay or mixer that can't be traced. These sales are kept but given half weight (flag U).
• (e) the price is more than 3× or less than 0.33× the previous FF.

IMDSTR's own purchases are real trades and are kept. However, its relists at a multiple are not counted as an independent ask unless they are the cheapest fillable listing.

2. Applied to the sales I could reach

The strategy contract is IMDSeatStrategy, a proxy pointing to implementation 0x16D3…2A14. It has a function buyTargetNFT(value,…,expectedId,target), and the target is Seaport 1.6.

Block	Token	Price (ETH)	Venue / route	Wash flag
26,050,642	1533	1.87	Seaport 1.6, bought by IMDSTR buyTargetNFT	Clean (protocol buyer)
26,054,425	195, 1628	~1.8–2.07 each (batch; price per token not resolved)	Seaport fulfillAvailableAdvancedOrders, a sweep by 0xDa71…C13E worth 22.40 ETH	Clean; price not exact, so excluded from M_s
26,054,540	806	1.95	Seaport 1.6, bought by IMDSTR	Clean (protocol buyer)
26,054,761	804	1.78 WETH (1.7622 net + 0.0178 fee, my reading)	Seaport matchAdvancedOrders, a WETH offer accepted by the seller	Clean; this is a bid fill
26,054,948	147, 178, 453, 1630	9.967 ETH total (~2.49 each, probably includes relay and bridge costs)	Relay solver → RelayRouterV3 multicall (cross-chain fill) to 0x1B45…d748	U (untraceable cross-chain funding); price unclear, excluded
26,054,951	1621	2.00	Seaport fulfillBasicOrder, bought by 0x1B45…d748	Clean
26,054,958	184	2.00	Relay multicall, 0xBe3d…7Daa → morello.eth	U
26,055,041	873	2.00	Relay multicall, 0xBe3d…7Daa → 0x1B45…d748	U
26,055,050	687	2.00	Relay multicall, 0xBe3d…7Daa → 0x92e9 (morello.eth)	U
26,055,373	29	1.99	Seaport fulfillBasicOrder	Clean
26,055,554	1035	1.98	Seaport fulfillBasicOrder	Clean

Not sales (transfers with 0 ETH, excluded):

• 827 at block 26,054,893
• 1895 at 26,054,846
• 809 at 26,055,615
• 888 at 26,055,713
• 743 and 868 at 26,055,887, sent through OpenSea TransferHelper bulkTransfer

Wash-trade findings:

• No confirmed wash trades. In this sample I saw no seller who is also the buyer and no token returning to a previous owner.

• Worth watching:

  • Seller 0xBe3d…7Daa made three 2.00 ETH Relay fills within about 18 minutes.
  • Buyer 0x1B45…d748 bought at least 6 tokens.

  I could not trace funding across chains, so these stay at U, half weight.

Figures:

Figure	Value	Block read	Basis
Naive floor	1.99 ETH	≈26,055,887 (OpenSea page fetched the same session; not pinned exactly)	OpenSea displayed floor. A search snippet an hour earlier said 1.90 ETH, which shows how much the displayed figure moves.
A₁ / A₃ / A₅	≈1.98 / ≈2.00 / ≈2.00 (inferred, not read)	26,055,554	I could not read the listing book. These come from the latest fills of listings: 1.98, 1.99, 2.00 ×4.
M_s	1.99	26,055,554	Clean sales at 1.87, 1.95, 1.98, 1.99 and 2.00 at full weight, plus the U-flagged 2.00 fills at half weight, give a weighted median of about 1.99–2.00. The 1.78 bid fill is kept at the low end.
Best bid B	1.78 WETH (last bid fill, block 26,054,761). IMDSTR's usable bid is 0 until the contract is topped up.	26,054,761 / ≈26,055,887	The strategy holds 0.598 ETH, which is less than the ~1.9 ETH it would need to buy one.
Fair floor	≈1.99 ETH (0.5·2.00 + 0.5·1.99 = 1.995; capped by A₃ ≈ 2.00)	≈26,055,887	Likely range 1.95–2.00
Spread	A₁ − B ≈ 1.98 − 1.78 = 0.20 ETH (~10.6%), or 0.21 ETH (11.1%) against the naive floor	mixed blocks

3. Uncertainty and data I couldn't access

• The 14-day sale list is incomplete.
  • OpenSea's events API rejected the request as unauthorized (HTTP 401).
  • Magic Eden's sales and activity API returned a bad-request error (HTTP 400).
  • The OpenSea and Etherscan pages came back without their data rows.
• Most prices came from a summary tool that reads Blockscout transactions. The decimal places for the WETH fill on token 804 had to be reconstructed. Prices for the batch transactions (0x435e…, 0x1737…) are not broken down per token.
• Listing book and top collection offer: not read directly. A₁, A₃ and A₅ are estimated from recent fills of listings, and B is the last filled bid, not a live bid.
• IMDSTR: I could not read priceMultiplier or getMaxPriceForBuy. The implementation source I retrieved does not appear to include the buying logic, which is probably in an imported NFTStrategy.sol. So the relist multiple is unknown.
• Block pinning: block numbers for the marketplace figures are approximate. They are the latest block I saw in the same session, not the block at which each page was read.
• Wash-trade checks: no buyer-funding trace was run (rule c). The U flags depend on how much Relay routing reflects genuine cross-chain buyers.

My overall confidence in FF ≈ 1.99 ETH is moderate, and it could fall within ±0.05 ETH. The bid side (1.78) is less certain than the ask side.

Sources: Etherscan: token · Etherscan: strategy contract · OpenSea · Blockscout API (transactions and token transfers) · imd-writeup

---

Sources offered but not confirmed by this machine:

• https://etherscan.io/address/0x0000198C940D8cD70Cb9ACeC5E3af8216ac57d2F — did not return a page

identity.md: a manipulation-resistant "fair floor"

Short answer: the fair floor is about 1.98 ETH, pinned to block 26055914 (25 Sep 2026, around 17:20 UTC). The naive floor is 1.99 ETH. The only bid I could see is a 1.78 WETH offer that was accepted at block 26054761. Using that bid, the spread is about 0.22 ETH (roughly 11%).

Main gap: I could only price about 5 hours of sales, not the full 14 days. The volume was too high for the fetch tool I had, and I couldn't read the live listings or offers directly (details in section 3).

1. Method

Step 1 — Clean the sales. Start with every identity.md transfer in the window. Keep it as a sale only if the buyer paid ETH or WETH to the seller in the same transaction (through Seaport, Blur, a Relay-routed Seaport fill, or a strategy contract's buyTargetNFT). Transfers where nothing was paid (plain safeTransferFrom, bulk transfers) are not sales. Where one transaction buys several items, price each item from its seller's payout plus the marketplace fee.

Step 2 — Flag wash trades. Drop a sale from the sales median if any of these hold:

• W1: buyer and seller are the same address.
• W2: the same token moves A→B→A, or loops through three wallets, within 7 days.
• W3: the buyer's funding path leads back to the seller within 7 days (direct transfer, same exchange deposit address, or same deployer or funder).
• W4: the price is more than 3× or less than 0.33× the rolling median and the buyer had no prior history.
• W5: an on-chain buyer that anyone can trigger (like buyTargetNFT) buys from a seller who is linked by funding to the address that triggered it.

Flips (bought and resold within 1 hour) are marked for review but kept.

Step 3 — Three inputs, read at block h:

• A (depth-weighted ask): the average of the 5 cheapest valid listings. A listing is valid only if the lister still owns the token, the marketplace approval is still live, the price is in ETH or WETH, and it isn't about to expire. No more than 2 listings per lister count, which stops one wallet from flooding the bottom. Stale listings that can still be bought stay in, because they are real supply. Listings that can't be filled drop out.
• S (sales median): the median price of the cleaned sales from the last 72 hours. If there are fewer than 10, widen the window up to 14 days.
• B (best credible bid): the highest of:
  • collection offers where the bidder's WETH balance and approval cover the price;
  • token offers on items priced at or below A;
  • the strategy contract's effective bid. This is its ETH balance, and it only counts if that balance is at least the cheapest listing. It can't pay more than it holds, and it buys at the floor and relists at 1.2×.

Step 4 — Combine:

• Fair floor F = min(A, max(B, S)), i.e. the sales median held between the bid and the ask.
• Spread = A − B, also reported as (A − B) divided by the midpoint (A + B)/2.

Why this design:

• A single low "troll" listing moves A by at most one-fifth of its gap.
• Fake wash prints are removed from S, and the median ignores outliers anyway.
• Because F can't rise above A, wash prints can't push it above what you could actually buy for. Because it can't fall below B, a real bid holds it up.
• A strategy contract's buys are real demand, but they're circular. Its treasury comes from fees on trading its own token, and because anyone can trigger buyTargetNFT with an order they choose, a seller can sell straight into it. So its bid is capped at its balance, and W5 applies to it.

2. Applied to identity.md

IMDSTR contract. The address 0x0000198C…d2F is a proxy named "Identity.MD Strategy". Its code is an IMDSeatStrategy contract that inherits from NFTStrategy. The parent's source isn't published, so I rely on press coverage of the NFTStrategy design for the 1.2× relist. It has made exactly two NFT buys ever, both through Seaport with the 1% OpenSea fee:

• #1533 for 1.87 ETH at block 26050642
• #806 for 1.95 ETH at block 26054540, triggered by 0x7BD5…03ED

Its ETH balance was 0.598 ETH at block 26055769. It can't buy at the current floor, so right now its effective bid is 0.

Sales I verified: blocks 26054425–26055914 (about 5 hours)

"OS/Seaport" means Seaport 1.6 with the fee going to OpenSea's fee address. Relay rows are cross-chain purchases routed through RelayRouterV3; their price may include Relay's fee.

Block	Token	ETH	Venue	Wash flag
26054425	1283 + 63 (seller 0x288c)	1.8825 each on average (3.765 total)	OS/Seaport sweep, buyer 0xDa71	clean
26054425	1104 + 424 (seller 0xf155)	1.90 each on average (3.80 total)	same sweep	clean; this seller then sold #806 to IMDSTR (see note at W5)
26054425	id not read (0xAc02)	1.95	same sweep	clean
26054425	id not read (0xf38E)	1.964	same sweep	clean
26054425	id not read (0x30b4)	1.965	same sweep	clean
26054425	2 ids not read (0xeafC)	1.975 each on average	same sweep	clean
26054425	id not read (0x1acf)	1.97	same sweep	clean
26054425	1628 (0x3bCe)	1.99	same sweep	clean
26054540	806	1.95	IMDSTR buyTargetNFT → Seaport	W5 check: caller ≠ seller; funding not traced
26054761	804	1.78 WETH	OS accepted offer (matchAdvancedOrders)	clean; this is a bid fill
26054937	804	1.98	OS/Seaport	review: resold 176 blocks (~35 min) after being bought at 1.78; counterparties differ; left out of S to be safe
26054946	1614	1.985 WETH	Relay	clean
26054948	147	1.99	Relay→Seaport, buyer 0x1B45	clean
26054948	453	1.995	Relay→Seaport	clean
26054948	178	1.997	Relay→Seaport	clean
26054948	1630	2.00	Relay→Seaport	clean
26054951	1621	2.00	OS/Seaport, buyer 0x1B45	clean
26054958	184	2.00	Relay, seller 0xBe3d	clean
26055041	873	2.00	Relay, seller 0xBe3d	clean
26055050	687	2.00	Relay, seller 0xBe3d	clean
26055373	29	1.99	OS/Seaport, seller 0x730a	clean
26055554	1035	1.98	OS/Seaport	clean
26055914	13	1.99	Relay→Seaport, seller 0x730a	clean
26055914	1073	2.00	Relay→Seaport	clean

Checking the per-item prices:

• Sweep at 26054425: OpenSea's fee was 0.21354 ETH, which is 1% of 21.354 ETH across 11 items. Sellers' payouts divided by 0.99 give the prices above. 1.0432 ETH of the 22.397 ETH sent was refunded.
• Relay batch at 26054948: the fee was 0.07982 ETH, 1% of 7.982 ETH across 4 items. One order didn't fill and 1.985 ETH was refunded.

Earlier sale outside the verified window: #1533, 1.87 ETH, IMDSTR, block 26050642. Clean.

Not sales (nothing paid, excluded): #1895 (26054846), #827 (26054893), #809 (26055615), #888 (26055713), #743 and #868 (26055887, bulk transfer).

Wash result: no same-address (W1) or round-trip (W2) trades in the verified window. I didn't trace funding for W3 or W5.

Results

Metric	Value	Pinned to
Naive floor (marketplace display)	1.99 ETH (OpenSea, via Alchemy); a separate OpenSea page read showed 2.0202	read at about block 26055914
A (depth-weighted ask)	about 2.00 ETH, estimated from listing fills at 1.98–2.00. I couldn't read the order book	26055914
S (cleaned median of the 26 sales above)	1.9825 ETH	window 26054425–26055914
B (best credible bid)	1.78 WETH (last accepted offer); IMDSTR's bid is 0 because its 0.598 ETH balance is below the floor	26054761 / 26055769
Fair floor F	min(2.00, max(1.78, 1.9825)) = about 1.98 ETH	26055914
Spread A − B	about 0.22 ETH (about 11.6% of the midpoint)	26055914 vs 26054761

3. Uncertainty and data I couldn't access

• Coverage: my table covers only blocks 26054425–26055914 (about 5 hours), plus the #1533 buy. OpenSea shows about $1.05M in 7-day volume, which likely means well over 100 sales in 14 days. I couldn't list them all: without shell access, each page of transfers returned only about 15 rows. So I have not produced one row per sale for the full 14 days. Running the same method over blocks ~25955000–26055914 would complete it.
• Order book: I got no live listings or offers. OpenSea's API returned 401, Magic Eden's returned 400, and the OpenSea pages load their data with JavaScript. So A is an estimate, and B is the last accepted offer rather than the current best offer. The spread could be narrower.
• Relay prices: the Relay rows' 2.00 ETH may include Relay's fee, so those prices could be slightly lower. Five token ids in the sweep weren't shown because the transfer list was cut off.
• IMDSTR's relist price: the 1.2× relist for IMDSTR is taken from the general NFTStrategy design, not from its code, which isn't published.
• Wash tracing: I didn't trace where buyers' funds came from, so W3 and W5 are untested.
• Data conflict: Etherscan shows 780 holders and Blockscout shows 1,349–1,352.