I keep coming back to one number when someone throws a small cap token in front of me and asks if it is real. Not the market cap, not the holder count, not the chart. I look at how much volume the thing reports against how much liquidity actually sits in the pool. That single ratio catches more nonsense than any other quick check I know, and it takes about ten seconds to eyeball once you know where the healthy bands are.
The idea is simple. Reported volume is a claim. Pooled liquidity is a fact you can read straight off the chain. When the claim is many multiples above the fact, something is manufacturing that activity, because a pool of a given depth can only support so much honest trading before slippage strangles it. When the claim is tiny against the fact, the pair is dead and you are looking at a position you may not be able to leave without eating a brutal spread. Both extremes are exits you want to see before you enter.
Why the pool is harder to lie about than the print
Centralized exchange volume is a print. It is a number reported by a venue, and the incentives to inflate it are old and well documented. Wash trading, self-crossing, market maker rebate farming, listings that pay for volume commitments. You cannot see the order book history that produced the number, so you are trusting the venue's honesty and its surveillance. For a small cap that mostly trades on second and third tier venues, that trust is worth very little.
DEX pool data is a different animal. The liquidity in an automated market maker pool is on-chain. You can read the reserves, you can read the swaps, you can reconstruct the fees. Faking volume through a DEX is not free, because every wash swap pays the pool fee and burns gas, and the price impact is deterministic. It is still fakeable, and I will get to how, but the cost floor is real and the data is auditable in a way a CEX tape never is. So when I build a turnover screen, I anchor it to the pool. Reserves for the denominator, honest swap volume for the numerator, both pulled from the chain.
The metric itself is just turnover. Take the token's trading volume over a window, typically a rolling twenty four hours, and divide it by the total value locked in its main liquidity pool or pools. That gives you a ratio. Volume of two against liquidity of one is a turnover of 2x. It sounds trivial, and it is, which is exactly why it belongs at the very front of a watchlist workflow before you spend real attention on anything.
What healthy actually looks like by tier
There is no single good number, because a mega cap and a microcap live in different physics. Deep liquidity naturally turns over more slowly relative to its size, and thin liquidity churns faster because a smaller pool serves proportionally more of the float. So I band the ratio by market cap tier rather than hunting for one magic threshold.
Rough bands I use as a starting point, and you should recalibrate these against your own universe rather than treating them as gospel:
- Large and mid caps. Turnover typically sits low, often well under 1x on a daily basis. Deep pools, patient flow. A large cap suddenly printing 5x turnover is either a real catalyst or a coordinated pump, and the chart usually tells you which within a day.
- Small caps. Healthy turnover tends to run somewhere in the low single digits, roughly the 0.5x to 3x zone. Enough activity to exit a normal position, not so much that the pool is obviously being churned.
- Microcaps. These run hotter and the bands widen, but once daily turnover climbs into the high single digits or double digits and stays there, I stop assuming organic interest. A pool that reports ten times its own depth in volume every day, day after day, is not being traded. It is being operated.
The two failure modes sit on either end. High turnover with shallow, static liquidity is the incentivized or fabricated signature. Somebody is running volume through a thin pool to juice a leaderboard, farm an emission, or paint a token as liquid ahead of a distribution. Near zero turnover against decent liquidity is the dead pair. The token exists, the pool exists, nobody is trading, and the moment you try to size out you discover the spread that the absence of flow was hiding.
Building it into a first filter
Here is the workflow I actually run when a name lands on the list. It is a gate, not a verdict. Passing it does not make a token good. Failing it saves you the hour you were about to waste.
- Pull the token's main pool or pools and read the reserves. Sum the total value locked across the venues that hold real depth, and ignore the dust pools with a few hundred dollars in them.
- Pull the rolling daily swap volume from those same pools. Use on-chain swaps, not an aggregator's blended CEX plus DEX figure, so your numerator and denominator describe the same market.
- Divide volume by liquidity to get turnover, then place it against the band for its market cap tier.
- If it sits far above the band, flag it as suspected manufactured activity and go look at who is doing the swapping before anything else.
- If it sits far below the band, flag it as a potential dead pair and mentally price in a much worse exit than the mid quote suggests.
- Only tokens that land inside the healthy band graduate to the real work, meaning holder distribution, unlock schedules, the team, the actual reason to care.
The one refinement worth adding early is a look at who generates the volume. Turnover can sit inside a healthy band and still be fake if three wallets are passing the same bag back and forth. So when a ratio looks suspiciously clean, count the distinct addresses behind the swaps and check whether volume collapses when you strip out the top two or three traders. Real markets have a long tail of participants. A wash operation usually does not, and that shows up fast once you look past the aggregate number.
Where the ratio lies to you
No single metric survives contact with people who are trying to beat it, and this one has known blind spots. Liquidity can be faked too, through pools that look deep but are single sided or dominated by the project's own treasury that can be pulled in a block. That inflates the denominator and makes turnover look calm on a token that is actually a rug in waiting, so a low ratio is never a safety certificate on its own.
The window matters as well. A genuine catalyst, a listing, an airdrop snapshot, a governance vote, can spike turnover for a day or two in a way that is completely organic. That is why I care about persistence. A one day spike is an event. Elevated turnover that holds for a week against flat, shallow liquidity is a pattern, and patterns are what the screen is really built to catch. Watch the trend, not the snapshot.
Treat the ratio as the cheapest first cut you can make, the thing that thins a fifty name watchlist down to the dozen worth reading properly. It will not tell you what to buy. It is very good at telling you what to skip, and on small caps most of the edge is in what you decline to touch.