Research
Counted NFT Transfers
NFTs on Ethereum follow a binary mobility paradigm: ERC-721 tokens can be transferred without limit, while Soulbound Tokens (ERC-5192) cannot be transferred at all.

- arXiv
- 2602.19199
- Published
- 2026-02-22
- Authors
- Qin Wang, Minfeng Qi, Guangsheng Yu, Shiping Chen
AI summary
Overview
Research area: Ethereum token standards and blockchain systems security (cs.CR), specifically non-fungible token (NFT) ownership mobility, with economic and security analysis drawn from DeFi markets.
Technical level: Intermediate. The paper assumes familiarity with ERC-721, Solidity hooks, gas costs, and basic DeFi concepts such as loan-to-value (LTV) ratios and wash trading, but it explains each mechanism it introduces.
Scope: The paper defines "counted NFT transfers," formalizes a transfer-bounded ownership model, specifies and implements the ERC-7634 extension to ERC-721, and evaluates its economic, security, and gas consequences through calibrated simulation.
What This Paper Is About
NFTs on Ethereum follow a binary mobility paradigm: ERC-721 tokens can be transferred without limit, while Soulbound Tokens (ERC-5192) cannot be transferred at all. The authors identify a gap between these extremes — no standard mechanism supports bounded transferability, where a token may change ownership only a finite, programmable number of times. The paper closes that gap with ERC-7634, which gives each token a transfer counter and a configurable cap L, and then studies what happens economically and security-wise when ownership mobility becomes a scarce, consumable resource.
Key Contributions
-
A formal model of counted transferability. The authors characterize the design gap between unrestricted ERC-721 transfers and permanently non-transferable SBTs, and formalize counted transfers as a transfer-aware state machine over the tuple
(k, L)— current transfer count and transfer limit — with mint, limit-update, transfer, and burn transitions plus stated safety and liveness properties. -
ERC-7634, a minimal backward-compatible standard. The extension defines three lightweight functions (
transferCountOf,setTransferLimit,transferLimitOf), two events (TransferCountIncreased,TransferLimitUpdated), and native-transfer hooks, requiring fewer than 60 additional lines of Solidity on top of OpenZeppelin v5's ERC-721_updatehook pattern. The proposal went through the standard five-phase Ethereum Improvement Proposal process (PR, Draft, Review, Last Call, Final) over nearly two years, from February 2024 to late 2025. -
Analytical characterization of behavioral, economic, and security effects. The paper derives a mobility premium and valuation model, a marginal mobility cost, a wash-trading deterrence argument, and a bound on recursive collateralization, plus a systematic treatment of the wrapper-bypass vulnerability and five (four enumerated) mitigation strategies.
-
Positioning and deployment guidance. ERC-7634 is placed within the broader NFT mobility design space alongside ERC-4907, ERC-5192, ERC-5484, ERC-5679, and ERC-6982, with practical cap-selection guidelines and a discussion of post-cap ownership outcomes: soulbound conversion, auto-burn, lock-and-release, and provenance freeze.
Main Findings
-
A mobility premium exists. Under a concave premium function
f(x) = x^γwithγ = 0.5, a token withL = 20and base valueV_base = 10ETH retains 7.07 ETH at the midpoint (k/L = 0.5), i.e., about 70.7% of base value, versus 5.00 under the linear model, 2.50 under the convex model (γ = 2), and 3.75 under the threshold model (τ = 0.2). Each remaining transfer therefore carries measurable value. -
Marginal transfer costs escalate. Under the concave model, the first transfer costs 10.6% of
V_baseatL = 5while the last costs 44.7%; atL = 50the same figures are 1.0% and 14.1%. Later transfers destroy disproportionately more value, creating endogenous friction on excessive trading. -
Moderate caps touch only a small share of tokens. In the calibrated simulation, a cap of
L = 10affects 8.8% of PFP tokens, 2.9% of art tokens, 14.2% of gaming assets, 0.8% of membership passes, and 5.5% of metaverse land tokens — fewer than 15% in every category. Transfer activity is highly right-skewed: median transfer counts are close to one everywhere, while P99 reaches 304 for gaming items and 83 for PFP collections. -
Cap design must be collection-specific. Estimated power-law exponents range from
α ≈ 1.8for gaming items (P99 = 304 transfers) toα ≈ 3.0for memberships (P99 = 9). A uniform cap ofL = 10constrains 0.8% of membership tokens but 14.2% of gaming assets. -
Wash trading becomes self-limiting under cap-aware pricing. With
α = 0.3inflation andV_base = 10ETH, a no-cap attacker gains roughly 3.0 ETH. Under caps, wash trading becomes unprofitable after roughly five trades atL = 10and after only three trades atL = 5; the simulation table shows cap-aware profit of 1.62 ETH after one wash trade atL = 5but −1.79 ETH after three, and 0.86 ETH after three atL = 10versus −0.83 ETH after five. -
Recursive leverage is structurally bounded. Each re-hypothecation cycle requires at least two native transfers, so maximum depth is
d_max = floor(L/2)and maximum leverage is the truncated geometric sum(1 − LTV^(d_max+1))/(1 − LTV). AtLTV = 0.7andL = 10, leverage falls to about 2.94× versus the unbounded 3.33× (an 11.7% reduction); atL = 6it is about 2.53× (a 24.0% reduction); atL = 4it is 2.19× (a 34.2% reduction); atL = 50the reduction is 0.0%. -
The wrapper-bypass attack is the main security limitation. Because limits apply only to native ERC-721 transfers, an attacker can move a token into a wrapper contract (consuming one transfer) and then reassign control of the wrapper freely. Bypass cost is
G_deploy + G_deposit + n · G_wrapper_transfer, withG_deploy ≈ 450,000gas, roughly $40 at 30 gwei. -
ERC-7634 is a budgeting primitive, not an absolute enforcement guarantee. It exposes remaining mobility rather than preventing circumvention under all execution paths. Mitigations carry measurable overhead: recipient allowlist ~8,200 gas per transfer, soulbound wrapper detection ~12,400, ERC-6982 lockable integration ~15,600, and a transfer cooldown period ~5,100.
-
Overhead is modest. The abstract reports that additional gas overhead remains below 11% per transfer and that a moderate
L = 10affects fewer than 15% of tokens under representative transfer distributions.
Methodology in Plain English
The authors start by surveying existing Ethereum token standards and building a comparison table across seven standards (ERC-721, ERC-4907, ERC-5192, ERC-5484, ERC-5679, ERC-6982, ERC-7634) on dimensions such as transferability, count awareness, value decay, DeFi compatibility, and anti-speculation strength. From that survey they identify the missing middle ground and write a specification: a small interface plus two rules — block a transfer if a finite limit has been reached, and increment a counter after every genuine ownership transfer (mint and burn do not count because one address is zero).
They then build a reference implementation on OpenZeppelin v5's ERC-721 using the _update hook, which replaced the older _beforeTokenTransfer/_afterTokenTransfer pair, so that the guard check and the counter increment happen atomically in one execution path.
For the economics, they define a value function that discounts a token as its remaining transfer budget shrinks, instantiate four shapes of that function (linear, concave, convex, threshold), and derive marginal costs, wash-trading profit conditions, and leverage bounds analytically. Those models are then checked numerically on 50,000 synthetic NFTs spanning five collection categories — profile-picture collections (e.g., Bored Ape Yacht Club), generative art (e.g., Art Blocks), gaming assets, membership passes, and metaverse land — with transfer frequencies drawn from power-law distributions calibrated to published empirical statistics. Simulation parameters use V_base = 10 ETH and a per-transfer cost g = 0.005 ETH. Finally, they analyze the security surface by threat-modeling a wrapper contract and costing out mitigations in gas.
Why This Matters
Impact on research. The paper reframes NFT mobility as a controllable protocol dimension rather than a binary attribute, and it gives the community a formal state machine, a standardized interface, and a set of analytical tools (mobility premium, marginal mobility cost, truncated leverage series) that other researchers can extend or test. It also contributes a rare protocol-level, rather than detection-based, approach to wash trading, complementing existing graph-analytics and platform-policy countermeasures.
Real-world applications:
- Event ticketing: tickets that can be resold a limited number of times, mirroring physical tickets that lose validity after resale.
- Gaming assets: items with a bounded trading lifetime, limiting infinite flipping while permitting normal in-game circulation.
- Memberships and credentials: passes that permit a few legitimate ownership changes before settling into a permanently bound state via soulbound conversion or provenance freeze.
- DeFi collateral: NFTs whose transfer budget caps re-hypothecation depth, limiting recursive leverage chains that the paper compares to dynamics that amplified the 2008 financial crisis.
Industry relevance. ERC-7634 is a submitted Ethereum standard that preserves full backward compatibility, so existing marketplaces, wallets, and lending protocols can interact with counted tokens without modification. Its "budget versus enforcement" framing tells protocol designers exactly what guarantee they are buying and what additional layers (compliant wrappers, registries, allowlists) they need for regulated or high-value assets.
Future Directions
- Stronger enforcement without sacrificing composability. The wrapper bypass remains open; the paper's own mitigations each trade away permissionless interaction or add gas, leaving room for designs that provide strict guarantees while keeping ERC-721 interoperability.
- Collection-specific cap selection. Because a uniform
L = 10constrains 0.8% of membership tokens but 14.2% of gaming assets, the paper argues collection-specific parameterization is needed for balanced policy outcomes — but the concrete rules for choosing caps remain guidelines rather than a validated methodology. - Empirical validation beyond simulation. The evaluation rests on 50,000 synthetic tokens calibrated to published empirical statistics; behavior under real deployments, real buyer pricing of remaining transfer counts, and real wrapper usage is not yet measured.
- Post-cap lifecycle design. The paper raises but does not fully resolve what should happen when the budget is exhausted — soulbound conversion, auto-burn, lock-and-release, and provenance freeze are presented as options whose comparative merits invite further study.
Target Audience
Protocol designers and Solidity developers building NFT standards or extensions; DeFi researchers studying collateralization, leverage, and market manipulation; marketplace and wallet engineers assessing compatibility with new token standards; and graduate students or practitioners in blockchain systems security who want a worked example of taking a token standard from gap identification through EIP standardization to economic and security evaluation. Readers without Ethereum background will find the economic sections accessible but should expect to look up ERC-721 hook mechanics.
Authors’ abstract
Non-fungible tokens (NFTs) on Ethereum currently follow a binary mobility paradigm: ERC-721 enables unrestricted transfers, whereas SBTs (ERC-5192) prohibit transfers entirely. We identify a design gap in which no standard mechanism supports bounded transferability, where ownership mobility is allowed but limited to a finite number of programmable transfers. We study counted NFT transfers and introduce ERC-7634 as a minimal realization compatible with ERC-721. The design…
The full abstract on arXiv