A trader who wants leveraged exposure to an asset usually has two instruments available and treats the choice as a matter of what the interface shows first. It is not a cosmetic difference. The two contracts distribute their costs along completely different timelines, and the one that is cheaper depends almost entirely on how long you intend to hold.
A dated future is an agreement to settle at a specified date. It has a fixed life, a known end, and a price that reflects both the current level of the underlying and the market's view of holding it until that date.
A perpetual has no expiry. It is designed to track the underlying indefinitely, which requires a mechanism to keep it from drifting away, since nothing forces it to converge to anything. That mechanism is a periodic payment between the two sides, and it is the defining feature of the instrument.
Both give the same directional exposure with leverage, and a position of the same size behaves identically to a price move in either. Everything that differs between them is in the carrying cost and in what happens as time passes, which is precisely the part that is invisible on a chart.
The two contracts behave identically to a price move. Everything that separates them is what happens while you wait, which is the part a chart does not show.
An expiry forces the contract and the underlying to meet. On the settlement date the contract is worth what the underlying is worth, by definition, and any gap between them before that date has to close by then. This is convergence and it is a mechanical certainty rather than a tendency.
That certainty is what allows a dated future to trade away from the current price without being wrong. A contract priced above the underlying is not making a prediction, it is reflecting the cost and benefit of holding the position until settlement, and that difference is arithmetic rather than opinion.
It also means the cost of the position is embedded in the entry price rather than charged along the way. You pay for the carry when you open, in the price you accept, and nothing further is deducted while you hold. That single property is the deepest difference between the two instruments.
A perpetual has no settlement to pull it towards the underlying, so an explicit mechanism does the job. At regular intervals, one side pays the other an amount determined by how far the contract is trading from the reference price, in the direction that discourages the deviation.
When the contract trades above the reference, holders of long positions pay holders of short positions, which makes being long more expensive and being short more attractive until the gap closes. When it trades below, the flow reverses. Nobody is being penalised; a price is being maintained by making the deviation costly.
The consequence for a holder is that the cost arrives continuously rather than being fixed at entry. A position held through many intervals accumulates payments that were not knowable when it was opened, in a direction that depends on where the contract trades relative to the reference throughout.
A dated contract ends, so a trader who wants to maintain exposure beyond its expiry has to close it and open the next one. That operation is called a roll, and it is where the cost that seemed absent reappears.
Rolling means crossing two spreads, paying two sets of trading fees, and accepting whatever difference exists between the expiring contract and the next one. That difference is the carry for the new period, and it is paid up front in the same way the original was.
This is the accounting people miss when comparing the two instruments. A dated future looks free to hold and is not, it is prepaid; a perpetual looks expensive to hold and is charged continuously. Over a long horizon they are two ways of paying for the same thing, and which is cheaper is an empirical question rather than a structural one.
Three properties explain the concentration. There is no expiry to manage, which removes an operational task and a recurring decision. There is a single contract per asset rather than a curve of them, which concentrates liquidity instead of dividing it across dates. And the tracking is tight, so the instrument behaves like the underlying with leverage attached.
For a trader holding for hours or days, these advantages are decisive and the carrying cost is small enough to be secondary. The instrument was designed for exactly that use and it does it better than a dated contract would.
The concentration is self-reinforcing, which is worth naming. Volume attracts volume, so the perpetual stays the deepest market, which makes it the cheapest to trade, which attracts more volume. That is a genuine advantage for most participants and it is not evidence that the instrument is better for every horizon.
The case for a dated future strengthens as the holding period lengthens. Over months, the accumulated cost of an unfavourable funding regime on a perpetual is unbounded, while the cost of a dated position is fixed at entry and cannot get worse regardless of what happens to sentiment.
It is also the instrument for anything where the horizon itself is the point. A position intended to express a view about a specific future date, or one that has to be closed by a known time for a reason outside the market, matches a contract that ends when the reason does.
The third case is hedging with certainty about cost. A hedge whose price can drift with a funding rate is a hedge with an open-ended expense, and for anybody who needs to state a total cost in advance, a fixed carry is worth more than a marginally better entry.
On a perpetual the carrying cost is open-ended. On a dated contract it is fixed at entry. Over months, that difference stops being a detail.
Almost all crypto derivatives settle in cash rather than by delivering the asset, which means the contract closes against a reference price and the difference is credited or debited. Nobody receives anything but the money, and the reference price therefore decides the outcome entirely.
That reference is normally an index built from several sources rather than the last trade on any single venue. The construction is deliberate: an index is far harder to move than one market, and a settlement that depended on one price would be a settlement somebody could influence.
Reading the index methodology before holding a dated position through expiry is a small, unglamorous piece of preparation that occasionally matters a great deal. Which sources are included, how outliers are handled, and what happens if a source stops reporting are all specified, and they all become the settlement price at the moment they are needed.
Where several dated contracts exist, they are not equally traded. The nearest expiry usually carries most of the volume, the next one carries much less, and distant dates can be thin enough that the quoted price is not a price anybody can transact at in size.
This changes the arithmetic of a long-dated position. A contract that is theoretically ideal for a six-month view can cost more to enter and exit than the extra carry it saves, and the comparison has to be made against the depth that actually exists rather than the instrument that theoretically fits.
The practical check is to look at the resting depth on the specific contract before assuming it is usable. That takes a moment, it is visible in any order book, and it resolves more instrument-choice questions than any amount of reasoning about carry.
Four lines separate the two. The funding stream on a perpetual, which is continuous and unbounded. The embedded carry on a dated contract, which is fixed at entry. The roll cost, which applies only to the dated instrument and only if you extend. And the spread and depth difference, which usually favours the perpetual.
Trading fees do not differ structurally, and the published schedule applies to both. What differs is how often you pay them: a perpetual held for six months pays them twice, at entry and exit, while a dated position extended across the same period pays them at every roll.
Adding the four honestly for your actual holding period is the whole of the analysis. It takes a few minutes, the inputs are all published, and it replaces a choice usually made by whichever tab was open with one made on the numbers.
There is no threshold in days that settles it, and anybody offering one is compressing a calculation into a slogan. What determines the answer is the size of the funding stream relative to the embedded carry over your specific horizon, and both of those move.
What can be said generally is directional rather than numeric. Short horizons favour the perpetual because depth dominates and accumulated funding is small. Long horizons favour the dated contract because a fixed cost beats an open-ended one. The crossover is where the arithmetic says it is, on the day you run it.
The useful discipline is to run it at all. Most traders never compare the two for a position they are about to hold for months, and the cost of not comparing is paid quietly in a stream of small payments that never appear as a single line anybody reviews.
A dated future settles on a specified date, which forces it to converge with the underlying, so its carrying cost is embedded in the entry price. A perpetual never expires, so a periodic payment between the two sides keeps it tracking, and that cost arrives continuously instead of being fixed at entry.
Because it settles later. The difference reflects the cost and benefit of holding the position until settlement, not a prediction. Expiry forces the two to meet, so that gap must close by the settlement date, which makes it arithmetic rather than opinion.
Closing an expiring contract and opening the next one to maintain exposure. It means crossing two spreads, paying two sets of fees, and accepting the difference between the two contracts, which is the carry for the new period. It is where the cost that seemed absent on a dated contract reappears.
No expiry to manage, a single contract per asset instead of a curve that divides liquidity, and tight tracking of the underlying. For horizons of hours or days these advantages dominate and the carrying cost is secondary. The concentration is also self-reinforcing: depth attracts depth.
When the holding period is long, because the accumulated funding cost on a perpetual is open-ended while a dated carry is fixed at entry. Also when the horizon itself is the point, and when hedging with a cost that has to be stated in advance rather than left to drift.
The contract closes against a reference price and the difference is credited or debited, with nobody receiving the asset. That reference is normally an index built from several sources rather than one venue's last trade, because an index is much harder to move than a single market.
No. The nearest expiry usually carries most of the volume and distant dates can be thin enough that the quoted price is not transactable in size. A contract theoretically ideal for a long view can cost more to enter and exit than the carry it saves, so check the depth first.
No, and any threshold offered is a slogan compressing a calculation. The answer depends on the size of the funding stream relative to the embedded carry over your specific horizon, and both move. Short horizons favour the perpetual, long horizons the dated contract, and the crossover is wherever the arithmetic puts it that day.