ClickOptionsThe Rollover Test
One trade, not a year of them. A daily Bitcoin short on the second break of a two-year trendline, held toward a target 205 days away. $100 goes in on each side and nothing is ever added. Run three ways: everything into one contract at 25×, everything in at 10×, and a quarter at a time. The first two lose the lot. The third finishes at $237. The instrument did not decide that — the setting did.
One trade, $100 each side, three ways
Every row below starts with $100 and never adds a cent. The perp puts it up as margin. The no-liquidation side spends it on a contract, and at every settlement whatever the account is worth goes into the next one. If the account is empty there is nothing to roll into and the trade is over. That is the fair comparison, and it is the one thing most explanations of this instrument leave out.
| Position it held | Perp | Perp left | Contract | Contract left | Lowest it got | |
|---|---|---|---|---|---|---|
| 25× · everything in$100 buys one contract | $2,419 | force-closed 02 Dec, day 1 | $0 | settled to nothing 03 Dec | $0 | $-0.00 |
| 10× · everything inlevel further out, position smaller | $989 | force-closed 09 Dec, day 8 | $0 | never closed | $7.67 | $0.87 |
| 25× · $25 a contractthe rest stays in the account | $605 | force-closed 02 Dec, day 1 | $0 | never closed | $237 | $32.35 |
The perp column is the same instrument at the leverage named in each row, with $100 of isolated margin. It is force-closed the moment price touches its level, intraday. The contract is never force-closed — it only ever looks at the settlement price — but it can settle with nothing left in it, which ends the trade just as finally.
All three start at $100 on 01 Dec 2025. The dashed line is the starting balance. The perp is flat at zero from the day it was force-closed.
25×, everything in, and both gone in two days
Both sides hold $2,419 of short from 86,244.1. At 25× a 4% move against you is the entire $100 — on either instrument. What differs is what happens when price gets there.
| When | Price | Against the short | |
|---|---|---|---|
| Short filled | 01 Dec 2025 | 86,244.1 | — |
| Perp force-closed | 02 Dec 2025 15:00 UTC | 89,348.9 | +3.60% against |
| Level where the stake is gone | — | 89,693.9 | +4.00% against |
| Contract still worth, at that minute | 15:00 UTC | $9.68 | 17 hours left to run |
| Highest price in those hours | — | 93,931.1 | +8.91% against |
| Last time it came back under the level | 02 Dec 15:03 UTC | 89,693.9 | three minutes, then never again |
| Contract settles | 03 Dec 2025 08:00 UTC | 93,022.9 | +7.86% against |
The perp is gone at 15:00 on the first day, on an intraday wick, at 3.6% against. The contract is not closed. It rides the whole spike to 93,931.1 and at the minute the perp died it still had $9.68 in it. It then had 17 hours for price to come back under 89,693.9. It came back for three minutes and never again.
The floor worked, and it still cost you everything. At settlement the position was down $190 and the loss stopped at $100. That is the floor doing exactly its job. It is just not the job most people think it does: it caps what you can lose on a contract, it does not keep you in the trade. Both instruments lost the same $100 — the contract bought 17 hours of still being allowed to be right, and the market did not oblige.
3 · Act two10×, and why surviving is not winning
So drop the leverage on both sides. The level moves from 4% out to 10% out, and the position drops from $2,419 to $989. Note that the contract buys slightly less than the perp's $1,000 for the same $100 — that gap is what the protection costs, visible in the very first number.
The perp is force-closed on 09 Dec, day 8. The contract is never closed, not once in 205 days — and still finishes at $7.67, having bottomed at $0.87.
Why, and it is not what people assume
| The same $989 of short, three ways to carry it | Ends at |
|---|---|
| Hold that position and never re-size it | $419 |
| Re-size it at every settlement, charge nothing at all | $16.80 |
| Re-size it at every settlement, with the real costs | $7.67 |
The swings account for $402 of the shortfall and the costs for $9. Carrying the position across 104 contracts cost $16 in total. Even if the contract had been completely free it would have finished at $16.80.
The mechanism. Every settlement puts the whole balance into the next contract. So the position is always the balance times the leverage, which means a 10% move against you costs you everything you have, and every recovery afterwards compounds from a smaller base. We were short and price fell +32.2% — the drops were not the problem. The bounces were: day one +7.9% took $100 to $21, and the February and April bounces did the rest.
The clearest way to see it. In mid-February price was 66,208 and the account was worth $109.42, more than it started with. In June price was 58,443.9, far better for a short, and it was worth $7.67. Price finished better and the account finished worse. That is path dependence, and no instrument is immune to it.
4 · Act threeA quarter at a time
Same trade, same $100, same rule that nothing is ever added. One change: only $25 goes into each contract and the rest stays in the account. The position now sits near $605 the whole way down instead of shrinking every time the market bounces.
6 contracts expired worthless along the way and not one of them ended the trade, because there was always something left for the next one. Final: $237, never below $32.35.
How much to put in each time
| How much goes in each time | Share of the $100 | Ends at | Lowest it got | Contracts that expired worthless |
|---|---|---|---|---|
| $100 into each contract | 100% of the account | $0.00 | $-0.00 | 1 |
| $50 into each contract | 50% of the account | $0.00 | $0.00 | 2 |
| $33 into each contract | 33% of the account | $263 | $13.21 | 6 |
| $25 into each contract | 25% of the account | $237 | $32.35 | 6 |
| $15 into each contract | 15% of the account | $182 | $59.41 | 6 |
Read this as a cliff, not a curve. A hundred a time is zero and fifty a time is still zero, because a run of settlements can empty the account before the trade works. Below that it flattens out. We are not presenting $25 as an optimum — the whole row is here so you can see the shape, and the shape is the point.
5 · How each side was settledThe same short, priced two ways
One short, priced two ways, from the same $100.
Perpetual future
- $100 of isolated margin at the leverage named in the row. Force-closed the moment price touches its level at any point in the day, not at a settlement.
- Taker fee 0.045% per side; real Bybit funding at every 8-hour settlement the position was open. A short collects funding when the rate is positive, which it mostly was.
- Once it survives, it is a fixed position. It stops caring about bumps, wicks and recoveries.
No-liquidation contract
- $100 buys a contract that settles at 08:00 UTC. Nothing can close it before then. The most it can lose is what went into it.
- At each settlement the contract pays out whatever it is worth, and that goes into the next one. The level that would take the whole stake is re-set each time, at the same distance from the new price — so what ends a contract is a single move bigger than that distance, not a move measured from where you first got in.
- Because the level is re-set from the current price, putting more in does not move it. Size and the chance of being wiped are independent, which is not true of a perp, where adding size at the same margin pulls the liquidation closer.
- Costs: the spread and fee on each contract, plus what it costs to hold it to settlement, calibrated on live ClickOptions quotes.
The rule of thumb, part five
A perp that survives its level is a fixed position. A contract that rolls puts the whole balance back in every time, so bumps are what it lives on and what it dies of. The wick that saves you intraday is the same wick that bleeds you over months — same feature, opposite sign, and the only variable is how long you hold.
6 · What this does not showRead these before quoting a number
- This is one trade. Episodes 1 to 4 are a year of signals each; this is a single position chosen because it stress-tests the rollover. It is an illustration, not a sample.
- It is deliberately outside the window. This instrument is built for intraday out to about a week, which is where the volume is. A 205-day hold is far outside that, which is the point of the episode — but do not read the result as what it does inside its own window. Episodes 1, 3 and 4 are that window, and the no-liquidation side wins all three.
- A longer contract does not fix it. At the same effective leverage a 16-day contract finishes at $6.76 against the 2-day's $7.67, with 13 settlements instead of 104 and more cost, not less. Leverage first, contract length second.
- The prices are modelled. The perp side is real exchange data; the no-liquidation side is calibrated to live quotes. That calibration is built on short-dated contracts, so treat anything beyond about a week as indicative.
- A perp at the right leverage beat all three. At 7× it held to the target. The claim here is not that one instrument wins — it is that one of them needs a level chosen in advance that nobody can know, and the other needs a sizing decision you can make on the first day.