Key Takeaways
- What it is: How far an asset's tradeable inventory is split across separately addressable venues, measured from the position of a specific order rather than as a property of the token.
- How it is built: Enumerate every venue and every distinct contract, snapshot each at a fixed block, convert each to a depth curve, then aggregate only inside the set one transaction can reach.
- Core expression: Depth(delta) is the notional tradeable before the marginal price moves by delta. Stated value is what the venue holds at the current price.
- Best use: Estimating what a given order size can execute against, and where the rest of the visible liquidity is stranded.
- Main limitation: Stated value moves on price alone, depth can fall while stated value rises, and a snapshot is not a promise: providers can withdraw before the order lands.
- Practical rule: Never compare venues by stated value. Compare them by depth within a stated band, and state which venues a single transaction can touch.
Who This Guide Is For
Read this page if you have used a total value locked figure as a proxy for how easy an asset is to trade. The gap between the two quantities is often more than an order of magnitude.
Swap-volume mechanics, wash trading and how trade counts are constructed are covered on DEX volume and on-chain liquidity. Why the same asset ends up existing as several non-fungible representations in the first place is covered on bridge flows.
Educational content. Not individualized financial advice.
What Does Crypto Liquidity Fragmentation Measure?
It measures how far an asset's tradeable inventory is split across venues that have to be reached separately. How fragmented it looks is set by what a particular order's execution path can touch in one transaction. A 50,000 dollar order routed on one chain faces almost none. A 5,000,000 dollar order that must cross a bridge to reach half the inventory faces a great deal.
The axes liquidity splits along
- Chain. Inventory on another chain requires a route, and routes cost time and money.
- Representation. Where the asset arrived by bridge, each bridge mints its own contract. Two representations on one chain are two assets that happen to share a name.
- Venue type. Automated market makers, order books and request-for-quote systems hold inventory in structurally different forms, and only the first has anything resembling a stated value.
- Pool and fee tier. The same pair on the same venue can run several pools at different fee levels, each with its own reserves and its own price.
- Price range inside a pool. In a concentrated design, inventory is placed over a chosen interval and is inert outside it.
- Time. Liquidity supplied only for the block containing a specific trade appears in no snapshot taken before or after it.
Plain-language definition
How spread out an asset's tradeable inventory is, and how much of it your particular order is able to touch.
Technical definition
For a pool at block h, stated value is the market value of both reserves at the current price. Depth is a function rather than a scalar: for a price move of size delta, it is the notional that can trade before the marginal price has moved that far.
How Is Crypto Liquidity Fragmentation Constructed?
A defensible measurement is built in six steps, and the fourth is the one most published figures skip.
- Enumerate every venue holding the asset, and every distinct token contract representing it. Grouping two bridge representations under one ticker at this stage guarantees an overstatement later.
- Snapshot at a fixed block per chain and record the heights. Chains have different block times, so a snapshot taken at one wall-clock instant sits at different points in each chain's history, and the skew is systematic rather than random.
- Convert each venue's state to a depth curve, not a value. For an automated market maker, evaluate the price-impact function against its reserves and range positions. For an order book, sum resting size by price band. An order book has no total value locked at all.
- Define the reachability set. Which venues can one transaction touch atomically, which require a bridge, and what is the latency of that bridge relative to the horizon of the order.
- Aggregate inside the reachability set only, and report everything else separately as stranded depth rather than folding it into a total.
- Account for the fee. Crossing a 1 percent fee tier costs 1 percent, so depth in that tier is only relevant to an order large enough that the extra size beats the fee. A router that ignores a well funded tier is usually right to.
Formula and Measurement Logic
Depth(delta) = notional tradeable before the marginal price moves by delta. Stated value = market value of inventory held, at the current price.
The relationship can be derived exactly for the simplest case, and that derivation is the benchmark everything else should be read against. In a constant-product pool holding reserves x and y whose product is held constant, the marginal price is y divided by x. Moving that price by a factor m requires supplying an amount of y equal to (sqrt(m) - 1) times y. Because the pool's total stated value equals 2y at the current price, the share of stated value consumed by that move is:
share of stated value = (sqrt(m) - 1) / 2
| Price move | m | Share of a constant-product pool's stated value consumed, one side |
|---|---|---|
| 1% | 1.01 | 0.25% |
| 5% | 1.05 | 1.23% |
| 20% | 1.20 | 4.77% |
| 100% | 2.00 | 20.71% |
A constant-product pool with no range choices at all offers under 0.25 percent of its stated value inside a 1 percent move. Adding the two sides of the last row, roughly 59 percent of such a pool's stated value backs prices outside a band running from half the current price to double it. Nothing is wrong with the pool; that is simply where a constant-product curve puts inventory.
Concentrated designs let a provider move that value into a chosen interval. Placed narrowly, near-mid depth rises by a large multiple. Placed widely, it falls. The stated value is identical in both cases, which is exactly why the stated value cannot answer the question.
How Should Crypto Liquidity Fragmentation Be Interpreted?
The only defensible summary is a depth curve at a stated order size, from a stated origin chain, at a stated block. Every scalar alternative discards the shape, and the shape is the answer.
Three ways stated value and depth move apart
- Providers widening ranges. During volatility, concentrated providers widen their intervals to reduce the risk of going out of range. Stated value holds or rises while depth at the touch falls.
- Price alone. A pool holding the volatile side of a pair is worth more when that asset appreciates, with no provider having done anything. Stated value in dollars is partly a price series wearing a liquidity label.
- Range exit. In a concentrated design, a position whose interval no longer contains the current price stops offering depth on one side entirely, while still contributing its full holding to the stated value.
Depth is a snapshot, not a commitment
Nothing obliges a liquidity provider to still be there when an order arrives. Positions can be withdrawn in the block before a trade, and inventory can be added for a single block and removed immediately after, so the depth a large trade meets can differ from a snapshot in either direction. Ordering within a block decides what a trade meets, a mechanism the Ethereum.org documentation on maximal extractable value describes in detail.
Step-by-Step Workflow
- State the order size and the origin chain first, because the answer depends on both.
- List every venue and every distinct contract, keeping bridge representations separate.
- Fix a block per chain and record the heights.
- Compute depth within a stated band for each venue, not the stated value.
- Express that depth as a multiple of 0.25 percent of the venue's stated value.
- Split venues into what one transaction can reach atomically and what it cannot.
- Subtract the fee tier cost before treating a tier's depth as usable.
- Sum depth only inside the reachable set, and report stranded depth separately.
- Compare that reachable figure against the order size, not against the total.
- Sample depth over several blocks to see whether it is stable or intermittent.
Worked Hypothetical Scenario
This example is hypothetical. Every figure is invented so the arithmetic can be verified.
One concentrated pool, stated value against depth
A concentrated-liquidity pool quotes a token against a dollar stablecoin at a mid price of 20.00 dollars and reports 50,000,000 dollars of stated value, broken down below by how far each provider position sits from the mid price.
| Distance of the position from mid | Stated value in that band | Share of pool |
|---|---|---|
| Within 1% | 1,900,000 | 3.8% |
| 1% to 5% | 6,100,000 | 12.2% |
| 5% to 20% | 12,000,000 | 24.0% |
| 20% to 100% | 18,000,000 | 36.0% |
| Beyond 100%, or currently out of range | 12,000,000 | 24.0% |
| Total | 50,000,000 | 100% |
Only 1,900,000 dollars, or 3.8 percent of the pool, sits inside 1 percent of the mid. A buy consumes only the inventory held above the mid, so if positions are roughly symmetric a buyer can take about 950,000 dollars before pushing the price 1 percent, which is 1.9 percent of the stated value. Stated value exceeds that usable figure by a factor of 52.6.
Against the constant-product benchmark this pool is doing well, not badly. A plain constant-product pool of the same 50,000,000 dollars would offer about 124,700 dollars inside a 1 percent move, so 950,000 is roughly 7.6 times the benchmark. Concentration bought a genuine improvement, and the headline number still overstated executable depth by more than fifty times.
Five venues, one asset
Now spread the same 50,000,000 dollars of stated value across the five places this asset actually trades. Three are pools on the origin chain at different fee tiers. Two are on other chains, each holding a representation minted by a different bridge, so the two are not fungible with each other.
| Venue | Stated value | Depth within 1%, one side | Depth as a multiple of the constant-product benchmark | Reachable in one transaction from Chain 1? |
|---|---|---|---|---|
| Chain 1 pool, 0.05% fee tier | 18,000,000 | 620,000 | 13.8x | Yes |
| Chain 1 pool, 0.30% fee tier | 14,000,000 | 180,000 | 5.2x | Yes |
| Chain 1 pool, 1.00% fee tier | 4,000,000 | 25,000 | 2.5x | Yes |
| Chain 2 pool, bridge A representation | 9,000,000 | 310,000 | 13.8x | No, requires a bridge |
| Chain 3 pool, bridge B representation | 5,000,000 | 95,000 | 7.6x | No, requires a bridge |
| Total | 50,000,000 | 1,230,000 | 825,000 reachable |
Stated value across every venue is 50,000,000 dollars. Total depth inside 1 percent of mid, everywhere, is 1,230,000 dollars, or 2.46 percent of that. Depth a single transaction originating on Chain 1 can route into is 825,000 dollars, or 1.65 percent. The 405,000 dollars sitting on Chain 2 and Chain 3 is 32.9 percent of all the near-mid depth in existence for this asset, and it is unavailable to an order that has to settle now.
An order of 2,000,000 dollars exceeds every near-mid figure here. It is 2.42 times the reachable depth inside 1 percent, so its realized price is set by the shape of liquidity further out, by the fee tiers it has to cross, and by whether a solver is willing to carry cross-chain inventory risk on its behalf. The 50,000,000 dollar headline contributed nothing to answering that.
What Can Make the Interpretation Wrong?
- Summing stated value across venues. The sum is arithmetically valid and describes nothing an order can do, because reachability and placement are both discarded.
- Treating two bridge representations as one asset. They are separate contracts in separate pools. Merging them overstates what a trade can reach and hides that a failure at one bridge impairs only its own version.
- Ignoring the price component. Stated value in dollars rises when the volatile side of a pair appreciates, with no provider having acted.
- Ignoring the fee tier. Depth behind a 1 percent fee costs 1 percent to reach, so it is inert for orders too small to justify it.
- Mixing venue types. Order books have no stated value at all, so any total that includes them is undefined rather than merely imprecise.
- Snapshot timing skew. Different block times mean a multi-chain snapshot at one wall-clock moment is not simultaneous, and the skew is systematic.
- Liquidity that exists for one block. Inventory supplied for the block containing a trade and withdrawn immediately after appears in no snapshot on either side of it.
- Treating depth as a promise. Providers can withdraw between the snapshot and the order, and a concentrated position that exits its range stops quoting one side entirely.
| Weak reading | What the data supports |
|---|---|
| This asset has 50 million dollars of liquidity | Venues holding it report 50 million dollars of inventory at the current price |
| Total value locked rose, so trading got cheaper | Held inventory is worth more; near-mid depth may have fallen at the same time |
| The asset trades on five venues, so it is well supported | Two of the five are unreachable inside the horizon of a settling order |
| The 1 percent fee tier has 4 million dollars behind it | Reaching it costs 1 percent, so it only helps orders large enough to justify that |
| Depth was 800,000 dollars, so my order will fill there | Depth is a snapshot; providers can withdraw before the order lands |
Cross-Network and Provider Comparison
Depth on another chain is chain-local unless a route exists, and the binding constraint is almost always latency rather than existence. A canonical rollup bridge with a mandatory withdrawal delay puts the far side outside the horizon of essentially every trade: that inventory is real, visible, and unavailable.
Intent and solver systems change the picture without removing the constraint. A solver can fill an order locally and carry the cross-chain rebalancing risk itself, making remote depth economically accessible. The price the trader receives then includes the solver's inventory and bridging cost, so it is not the remote venue's quote, and the gap widens precisely when the route is congested.
Transaction cost sets a floor on which venues are worth routing into. On an expensive chain, splitting an order across a small pool can cost more in gas than the price improvement it buys, so that pool is unreachable in practice for small orders even though it sits on the same chain. Cheap chains invert this: routing across many tiny pools is affordable, so the effective venue count is much larger and a snapshot covering only the biggest pools understates depth.
Providers disagree on which pools qualify, whether order-book venues count at all, and how representations of the same asset are grouped. Those three choices produce most of the variance in published liquidity figures, well ahead of any difference in the underlying chain data.
Advanced Analytical Methods
Aggregated depth curves
Combining venues correctly means adding their depth curves, not their stated values. The combined curve is what a splitting router faces, and it is convex in a way a sum of scalars cannot represent.
Concentration multiples
Dividing measured near-mid depth by 0.25 percent of stated value produces a unit-free concentration multiple, comparable across pools of very different sizes and across venues using different curve designs.
Time-weighted depth
Sampling depth every block across a window and reporting the mean alongside the minimum separates stable inventory from liquidity that appears only around specific trades. The gap between the two is itself the measurement of interest.
Route-constrained aggregation
Recomputing the reachable set under several latency tolerances, from same-block to several minutes, shows how much of an asset's depth is a function of patience rather than of size.
Realized-fill measurement
Effective spread computed from actual executions rather than from pool state captures everything a snapshot misses, including ordering effects and withdrawn positions. It is backward looking, the price of being the only measure that reflects what happened.
Practical Checklist
- I stated the order size and the origin chain before computing anything.
- I kept each bridge representation as a separate asset.
- I measured depth within a stated band rather than reading stated value.
- I separated venues one transaction can reach from those it cannot.
- I reported stranded depth separately instead of folding it into a total.
- I recorded the block height used on each chain.
- I compared reachable depth against the order size, not against the total.
- I described depth as a snapshot rather than as a commitment.
Frequently Asked Questions
Is total value locked the same thing as liquidity?
No. Total value locked is a valuation of the inventory a venue currently holds, priced at the current market price. Depth is the notional that can trade before the marginal price moves by a stated amount. The first is a single number, the second is a curve, and collapsing the curve into the first number is exactly where the information is lost.
Why does a large pool have so little depth near the mid price?
Because liquidity has to be placed somewhere across the price line, and only the part placed close to the current price is available to a small move. In a plain constant-product pool the value is spread across every possible price, and the arithmetic gives just under 0.25 percent of stated value for a 1 percent move. A concentrated design lets providers move value into a narrow band, which raises near-mid depth sharply, or into a wide one, which lowers it.
Can I add up the liquidity for one asset across chains?
You can add the stated values, and the sum will not describe anything a trade can do. Depth on another chain is reachable only if a route exists inside the horizon of the order, and a bridge with a withdrawal queue puts it outside that horizon for most trades.
Does rising total value locked mean it got easier to trade?
Not reliably, and the two frequently move in opposite directions. Providers widening their price ranges during volatility raises the stated value they hold while lowering depth at the touch. Stated value also rises on price alone, because a pool holding the volatile side of a pair is worth more when that asset appreciates and nobody has acted. Checking depth near the mid separately is the only way to tell which happened.
Why do fee tiers split liquidity for the same pair?
Because each fee tier is a separate pool with its own reserves and its own price. Providers choose a tier based on how volatile they expect the pair to be, so the same pair can hold meaningful inventory in several tiers at once. Crossing a 1 percent tier costs 1 percent, so its depth only helps an order large enough that the extra size beats the fee.
How does a routing aggregator reduce the practical cost of fragmentation?
By splitting one order across several pools so that each takes only the portion it can absorb cheaply, which produces a better average price than routing the whole order to the deepest single venue. This addresses the execution consequence of fragmentation without addressing fragmentation itself: the liquidity remains scattered, and the aggregator pays gas on each leg, so the benefit shrinks for small orders where fixed costs dominate. It also fails during periods when a chain is congested and additional legs become expensive or unreliable.
What is the difference between fragmentation across venues and across price ranges?
Fragmentation across venues means the same pair trades in several separate pools, so any single one holds part of the total depth. Fragmentation across price ranges happens inside one pool, where providers have placed liquidity in different bands and only the band containing the current price is active. The first is addressed by routing across venues; the second cannot be, because the inactive liquidity is unavailable at any venue until the price moves into its range. Both reduce executable depth, for different reasons.
Why does a bridged representation trade at a different price from the canonical asset?
Because they are different instruments backed by different guarantees, even when they are meant to be interchangeable. The representation's value depends on the bridge's escrow remaining intact and the bridge remaining able to process redemptions, so any doubt about either widens the gap. Arbitrage closes it only to the extent someone is willing to hold the representation while a redemption completes, which takes time and carries that same risk. Persistent discounts on a bridged asset are a market assessment of the bridge, not a data error.
How should fragmentation be measured, and what does a venue concentration index miss?
A concentration index over venue shares captures how spread out the liquidity is, but it treats every unit of liquidity as equivalent. It misses that liquidity within a pool may sit far from the current price, that some venues are connected by aggregator routing and effectively act as one, and that liquidity on a chain a trader cannot reach cheaply is not available to them at all. A useful measure starts from executable depth at a given order size rather than from the distribution of headline totals.
References
These sources should be reviewed during editorial verification. They support data structures and methods, not the hypothetical conclusion. Provider formulas, chain rules, and APIs can change. Confirm current documentation before publication.
- DefiLlama: Data Definitions: How total value locked is defined, which is the quantity this page argues is not depth.
- DefiLlama: Methodology: How a cross-chain aggregator decides which venues and representations enter a total.
- Dune: Curated Data Overview: Normalized cross-chain venue datasets used to enumerate where an asset trades.
- Ethereum.org: Gas and Fees: Transaction cost mechanics that set the minimum viable order size per venue.
- Ethereum.org: Maximal Extractable Value: Transaction ordering and searcher behavior, which determine what a trade meets rather than what a snapshot showed.