A buy signal just fired on an altcoin. Clean setup, everything lines up. But the token's unlock calendar says a large tranche of supply hits the market in nine days. Do you take the trade?

Most traders answer with a gut feeling. This post is about turning that gut feeling into something you can test against history — and about the one trap that quietly makes most such tests worthless.

We'll start where the argument is loudest right now: Hyperliquid.

The Hyperliquid question: is FDV even the right number?

As of late May 2026, HYPE trades around $65. Its circulating market cap sits near $16B, but its fully diluted valuation (FDV) — the value if every token were already in circulation — is roughly $60B. Almost a 4x gap. That gap is the whole debate.

The analyst DeFi Monk argued recently that valuing Hyperliquid at its FDV is misleading, and that the real number sits much closer to the circulating cap. His core point is sound: FDV assumes 100% of supply hits the market today, at today's price. That never happens. Equities don't even work this way — nobody values Nvidia on the maximum number of shares it could ever authorize, only on shares actually outstanding plus realistic dilution.

About 39% of HYPE's supply is parked as "future emissions" with a schedule the team hasn't fully disclosed. His move is to exclude that chunk from any real market-cap discussion, because nobody knows when (or if) it ships — and because the team might deploy it accretively, funding buybacks rather than dumping.

Here's what makes that defensible for Hyperliquid specifically: it has a real demand sink. The protocol's Assistance Fund routes the overwhelming majority of trading fees — around 97% — into open-market HYPE buybacks. More than $1.16B has gone into purchases since launch, at an intensity of roughly 7% of market cap per year, four to five times that of Ethereum or BNB, and funded by actual fees rather than token issuance.

But notice the shape of the argument. "Ignore 40% of supply because the team will probably be accretive" is a trust bet. It holds only as long as fees stay high and the team keeps choosing buybacks. And buybacks are volume-sensitive — they shrink exactly when emission pressure would hurt most. Add the team's own 23.8% allocation vesting out toward 2027, and you have real, scheduled sell pressure regardless. The conclusion is reasonable; the reasoning is the kind that becomes dangerous the moment you copy-paste it onto a token that has no buyback engine underneath it.

What a high FDV/MCap ratio actually tells you

Strip away the specifics and FDV/MCap is just the dilution multiple: total supply divided by circulating supply. A high ratio means most of the supply isn't trading yet. It tells you how much future supply exists — not when it unlocks, to whom, or whether anything absorbs it. And those three things are what actually decide the outcome:

"Exit liquidity" is what happens when those line up badly. A tiny float pumps the price on thin liquidity; the huge paper FDV becomes the marketing number ("a $10B project!"); then insiders unlock into the retail demand that the FDV anchored. The low float isn't an accident — it's the setup.

So a high FDV/MCap ratio is a flag, not a verdict. Hyperliquid leans benign because of its buyback. Most tokens don't have that excuse.

From valuation to price action: what unlocks do

Valuation is the slow story. The fast story is what actually happens around an unlock date — and here the folklore is mostly wrong.

The popular belief is that price gets pumped before an unlock to bait exit liquidity. The aggregate data says the opposite. Across a study of more than 16,000 unlock events, large unlocks tend to drift down into the event: the market front-runs them roughly 30 days out, the decline accelerates in the final week, and price typically stabilizes within about 14 days afterward. Somewhere near 90% of unlock events show negative price action, and most of that move is priced in before the date, not after.

A few details matter more than raw unlock size:

• Size vs. liquidity, not size alone. A 1%-of-circulating unlock into a thin book hurts more than a larger one into deep liquidity. The number to watch is unlock value divided by average daily volume.
• Who receives it. Team and investor cliffs are the most destructive. Ecosystem unlocks can be neutral or even positive, because they often coincide with genuine usage growth.
• Cliff vs. linear. Linear vesting spreads sell pressure thin; cliffs concentrate it.

The orchestrated pre-unlock pump does exist — but as a manufactured exception in low-float setups, not as a pattern you can reliably trade. The honest, repeatable edge isn't riding the pump. It's avoiding the anticipatory bleed and re-entering after price stabilizes.

Turning it into a rule

This is the part you can systematize. Instead of asking "should I take this signal?" as a vibe, you make the unlock calendar a feature of every decision:

• A blackout: don't take long entries when a large unlock (say, above 1% of float, or a meaningful multiple of daily volume) falls within the next 7–30 days.
• A re-entry delay: wait until the unlock has passed and price has had its ~2-week stabilization window, then let your normal signal fire.

Note the reframe at the heart of this. "Inflation" for a token isn't the FDV/MCap ratio — that's a static snapshot. What hurts holders is the flow: annualized new sellable supply as a percentage of float, net of any sink, measured against demand. DeFi Monk's instinct at the end of his argument was right — the inflation that matters is staking rewards plus team unlocks. Track what actually reaches the market.

The hard part: testing it retroactively without fooling yourself

Now the real question: how do you check whether this rule would have helped, using history? This is where most attempts quietly break.

The unlock schedule is public and known in advance, so using "days to next unlock" as an input at decision time is not look-ahead bias. That part is fine. The trap is subtler. Unlock data providers give you today's schedule. But schedules get changed, accelerated, delayed; tokens get burned. If you take today's revised calendar and apply it to a bar from eight months ago, your backtest now "knows" about schedule changes that weren't announced yet. That is look-ahead bias wearing a disguise — and it inflates results in exactly the direction that makes a useless rule look clever.

The only honest fix is point-in-time data: at any historical bar, you must use the schedule as it was known on that day, not the cleaned-up version you can see now. Practically, that means snapshotting the forward schedule every day and querying the snapshot that existed at the bar's date. For periods before your snapshots begin, the honest default is to disable the feature and say so — not to silently backfill today's schedule.

Two more disciplines decide whether the result is evidence or anecdote:

• Survivorship. Tokens that died are missing from most unlock datasets. If you only test the survivors, you flatter the strategy. Acknowledge the gap.
• Sample size. A filter removes trades. If applying it drops you below ~30 trades, you don't have a result — you have an anecdote. Under 30 trades is not evidence, no matter how good the equity curve looks.

What you'd actually need to build it

Here's the unfun truth: the unlock calendar is the easy part — it's public, and several providers serve it. The hard part is a backtesting engine that is genuinely disciplined about point-in-time data, that never lets a forward-looking value leak into a past decision, and that lets you measure a single filter's effect honestly against a clean baseline.

That infrastructure — look-ahead-checked, point-in-time, with filter-effect measurement — is exactly what Backtesting Arena is built on, and it's reachable over our API (REST, MCP, and pay-per-call). If you want to build unlock-aware testing yourself, you're not starting from zero on the part that's actually hard. You're bolting an unlock feed onto a scaffold that already treats history honestly. The rest is just deciding your thresholds — and being willing to delete a rule that only looks good because it stopped trading.

Study the Past — Improve your Future. 🥋