Reading depth properly: the ladder, not the headline
Depth is the amount that can be traded before the price has moved a stated distance. It is the only liquidity number that answers a question a trader actually has, and it is almost never the number shown first. This entry builds a depth ladder from reserves, explains why the buy and sell sides rarely match, and shows what a concentrated range does to a ladder that looked solid a moment earlier.
- Question
- What is the correct way to measure how liquid a Solana pair is, given that the headline liquidity figure does not answer the question a trader is asking?
- Balances used
- Both reserve amounts, the fee parameter, the liquidity distribution across intervals where the venue has one, and the set of venues quoting the pair.
- Out of reach
- Whether depth will still be there in an hour, and whether the providers behind it intend to keep it in place.
- What would overturn it
- A pool where trades of increasing size fill at a constant average price, which no bounded reserve pool can produce.
- Confidence
- Firm on how a ladder is computed. Provisional on any inference about why a ladder is shaped the way it is.
The short answer
Depth is the size that can be traded before the price has moved a stated distance. Written properly it always carries three qualifiers: how far, which direction, and on which venue. Depth at one per cent on the sell side of a named pool is a statement that means something. The word liquidity on its own is not.
The reason this matters is that depth is the number that decides execution cost, and execution cost is what a position actually pays. A pair can hold a large amount of capital and still be expensive to trade if that capital is placed where the price is not, which is why the headline figure and the trading experience so often disagree.
A ladder is simply that measurement repeated at several distances. Read together, the rungs describe the shape of the market rather than its size, and shape is where the useful information lives.
Why the headline figure fails
Total value locked adds up what a pool holds and converts it to a common unit. That is a real quantity and it does bound the market: a pool cannot support depth it does not have capital for. The failure is that it says nothing about arrangement, and arrangement is decisive on any venue that lets providers choose a range.
Consider two pools holding identical capital. One spreads it across the entire price range, so depth is modest everywhere and falls away gradually. The other concentrates it in a narrow band around the current price, so depth at half a per cent is enormous and depth at three per cent is close to nothing. Both report the same headline. A trader sizing an exit meets two completely different markets.
There is a second failure that is easier to fix and just as common. Value locked is denominated, usually in a major asset, so it moves when that asset moves even though no liquidity has changed hands. A pool whose headline figure fell overnight may have lost providers or may simply be denominated in something that fell. Reading raw balances rather than the converted total removes that ambiguity entirely.
Building a depth ladder
For a classic constant-product pool the ladder can be computed in closed form from the two reserves. To find the size that moves the price by a fraction, solve the invariant for the input amount that produces the target ratio. No sampling, no simulation, no data provider: two balances and some algebra give an exact answer.
For a concentrated venue the computation walks the liquidity distribution instead, using the interval and tick structure described in Orca's developer documentation. The price moves through intervals, and the amount available in each interval is consumed before the price enters the next one. The result is a piecewise calculation rather than a formula, and it produces ladders with sharp cliffs where a crowded interval ends.
In both cases the ladder should be built per side. The buy side asks how much of the quote asset can be spent before the price rises by the stated distance; the sell side asks how much of the base asset can be sold before it falls by the same distance. Reporting one number for both is a simplification that hides the asymmetry that matters most.
Depth on a bonding curve is one-sided
A token still on a launch curve has no second reserve in the ordinary sense. The curve is a function that sets a price from supply, and depth in the sell direction is bounded by whatever has been paid in so far. That makes the buy and sell ladders genuinely different objects rather than two sides of one symmetric calculation.
The practical result is that curve depth grows as the curve fills, and the shape of that growth is fixed by the curve parameters rather than by anyone's decision to provide capital. A reader carrying pool intuitions across without adjusting them will overestimate what an early curve can absorb and underestimate how quickly the price moves on ordinary size.
When a token migrates from a curve to a pool, the ladder changes shape at that moment even if the price does not. That discontinuity is a structural event rather than a market one, and treating the two ladders as a continuous series produces a step change that never happened in the market itself.
A worked ladder, illustrative
Two pools with the same headline
Illustrative figuresAll figures are invented to expose the shape of the problem and describe no real pool. Pool Wide holds 2,000 units of token A and 200,000 units of token B spread across the full range. Pool Tight holds the same amounts but concentrated so that most of the capital sits within roughly one and a half per cent of the current price. Both quote 100 B per A and both report the same headline capital.
| Distance from mid | Pool Wide absorbs | Pool Tight absorbs | What the pair of numbers shows |
|---|---|---|---|
| 0.5% | 5 | 34 | Concentration wins decisively for ordinary size |
| 1.0% | 10 | 62 | The advantage persists while inside the band |
| 2.0% | 20 | 78 | Tight is running out; the curve flattens |
| 5.0% | 51 | 84 | Wide keeps supplying; Tight has almost stopped |
| 10.0% | 105 | 88 | The ranking reverses for a real exit |
The reversal in the last row is the entire argument for reading a ladder rather than a number. For routine trading Pool Tight is far better. For an exit of meaningful size it is worse, and the headline figure that both pools share would have told a reader neither of those things.
Why the two sides differ
Asymmetry has two ordinary causes and neither implies anything sinister. The first is arithmetic: on a constant-product curve the cost of moving the price a given percentage is not symmetric in the two directions once fees are included, though for small distances the difference is minor.
The second cause is placement, and it dominates on concentrated venues. Providers frequently place ranges above or below the current price deliberately, either because they are willing to sell into strength, or because they are accumulating, or simply because they set a range once and the price has since moved. The resulting ladder is lopsided, and the lopsidedness is a fact about provider positioning rather than about demand.
For a reader the useful discipline is to record both sides every time and to resist explaining the gap. A pair that is cheap to buy and expensive to sell is a real and important observation. Why it is that way is a question about people, and the pool does not contain the answer.
What ranges do to a ladder
A concentrated position quotes only inside its interval. When the price crosses a boundary the position has been fully converted into one asset and stops contributing depth entirely. Because boundaries must land on the discrete steps a venue defines, and Meteora's protocol documentation sets out one such bin scheme, these exits often cluster, and a ladder can fall off a cliff at a specific distance rather than tapering.
That produces a market behaviour worth naming: a pair can trade smoothly for hours and then move violently on an order no larger than several that preceded it. Nothing was withdrawn. The price simply reached the edge of where the liquidity had been placed, and the next interval was sparse.
Reading for this is straightforward. Build the ladder at more distances than feel necessary, look for the rung where the incremental absorption collapses, and treat that distance as the edge of the reliable market. Anything beyond it should be planned for as if the pool were much thinner, because at that point it is.
A ladder is a snapshot
Every rung above describes the pool as it stands right now. Providers can add, withdraw or reposition at any time, and on concentrated venues repositioning is routine rather than exceptional. A ladder is a measurement of the present, never a commitment about the next hour.
Sizing a trade against depth
The practical use of a ladder is to invert it. Instead of asking what a trade will cost, ask what size the pool supports at a cost you are willing to pay, and treat that as the maximum. This turns depth from a descriptive statistic into a constraint, which is how liquidity should be used.
The same inversion applies to anyone planning activity around a pair rather than through it. Sizing routed flow sensibly means comparing intended volume against the depth that exists, because volume pushed through a pool far thinner than the flow simply moves the price and reports a large number that describes the pool being shoved rather than the market trading. Practitioners on that side of the question work through how much volume a token needs in exactly the same units a ladder produces, which is why the two readings belong together.
For an exit the calculation is stricter, because an exit has to happen whether or not the price cooperates. The right question is what a full exit would meet across all reachable venues at once, not what a comfortable partial exit would meet on the venue with the nicest interface. A position that can only be closed by walking a ladder down five per cent is a different position from the one the entry price suggested.
Four measures compared
| Measure | What it measures | Good for | Weak because |
|---|---|---|---|
| Total value locked | Capital present in the pool | Bounding the market, coarse screening | Ignores arrangement and moves with denomination |
| Depth at a distance | Size tradable before a stated move | Sizing trades and planning exits | Snapshot only; must be stated per side and venue |
| Reported volume | Turnover over a period | Gauging activity, not capacity | Says nothing about what the pool can absorb now |
| Realised slippage | Cost of trades that already happened | Auditing execution after the fact | Backward looking and biased by who traded |
A depth-reading checklist
- State the distance, the side and the venue with every depth figure, or the figure is not a measurement.
- Compute both sides separately and record the asymmetry rather than averaging it away.
- Extend the ladder past your realistic exit size, not just past your comfortable trade size.
- On concentrated venues, look for the rung where incremental absorption collapses and treat it as the edge.
- Work from raw reserve balances rather than a converted headline so denomination moves do not contaminate the reading.
- Fix the venue set before measuring and keep it fixed, so a later reading is comparable with this one.
- Re-read the ladder before acting, because providers reposition and a stale ladder is worse than none.
Reading depth this way takes a few minutes and removes most of the surprises that come from trusting a single liquidity number. Desk workflows on the trading side converge on the same list, and an operator evaluating a professional Solana volume bot is asking the identical question in reverse: how much can this pool take before the price stops behaving.
What a ladder cannot tell you
A ladder is a measurement of capacity at an instant. It cannot tell you whether the capital will remain, because providers are free to withdraw or reposition without notice and frequently do. It cannot tell you whether the price is reasonable, because the pool has no view on value. And it cannot tell you whether anyone will trade, because depth describes supply of liquidity rather than demand for it.
It also cannot be turned into a return. Deep pools are not therefore profitable to provide to, and an LP position in a deep pool carries the same structural exposure as one in a thin pool: the basket rebalances towards the weaker asset, and the position can end up worth less than simply holding the two assets outright. Fee income may or may not cover that gap, and no ladder tells you which.
There is one more boundary worth naming. A ladder describes a pool, and a trader interacts with a router whose splitting behaviour is documented separately, as the Jupiter developer documentation sets out. If the aggregator splits an order across three venues, the realised cost reflects a combination of ladders rather than the one that was measured, and it can be better than any single pool would suggest. Measuring one pool and reasoning about the whole pair is a category error that flatters thin venues and understates deep ones.
Used within those limits the ladder is the most honest liquidity measurement available, because every rung is derived directly from balances that anyone can read. Used past them it becomes another confident number in a field that already has too many.
Questions this entry gets asked
What does depth at one per cent mean?
It is the size that can be traded in one direction before the pool price has moved one per cent away from where it started. It is quoted per side and per venue, so a complete statement names the distance, the direction and the pool. Quoted properly it answers the question a trader has, which is how much can be done here before the price stops cooperating.
Why is total value locked still quoted everywhere?
Because it is easy to compute and easy to compare across pools, and because it was the first liquidity metric that dashboards adopted. It is not useless: it bounds how much capital is present. It simply does not describe how that capital is arranged, and arrangement is what decides execution on any venue with concentrated ranges.
Should depth be summed across venues?
Only if a router can actually reach all of them and only if the sum is stated as such. Adding depth across pools that an aggregator will genuinely split across is a fair description of what a trader can use. Adding depth from a pool nothing routes through inflates the figure with liquidity that will never fill an order.
Why do the buy and sell sides of a ladder differ?
Because the two reserves are rarely equal in effective size relative to the trade being measured, and on concentrated venues because liquidity is placed asymmetrically around the current price. A pool can be much easier to buy into than to sell out of, and that asymmetry is the single most useful thing a ladder shows that a single liquidity figure hides.
Does a deep pool mean a safe token?
No. Depth is a statement about execution capacity and nothing else. A pool can be deep around an asset with no future, and a thin pool can sit under something perfectly sound. This desk treats depth strictly as a cost-of-trading measure and makes no claim about the quality of anything a pool happens to quote.
How far out should a ladder be read?
Far enough to cover the size you would actually need to exit. Reading only the half-per-cent band is comfortable and misleading, because the interesting failures happen further out where concentrated ranges run out. A ladder that stops before your realistic exit size has not answered your question.
Can depth be faked?
Depth itself is backed by real balances, so it cannot be conjured. What can mislead is where the balances sit: a large position placed in a range the price is unlikely to reach contributes to value locked while contributing nothing to tradable depth. Reading the distribution rather than the total is the defence, and it is available to anyone.
Filed in Pools by The Depth Ledger Desk. Every quantity inside a worked example on this page is invented for teaching and describes no real pool. Nothing here is advice about what to buy, sell or supply, and liquidity provision can end with a position worth less than holding the two assets. Terms used above are defined in the liquidity glossary.