There is a size below which a correct trade is a losing trade, and almost nobody works out what theirs is before they start. They work it out afterwards, from a run of small positions that were individually right and collectively down, and they usually blame the strategy rather than the ticket size.
The arithmetic is a single division. What makes it worth writing down is that on event markets, and on-chain ones in particular, the answer often collides with a second number from the opposite direction, and the two together rule out a whole class of contract that looks perfectly appealing on the screen.
Four costs, and only two of them scale
Sort every cost of a round trip into two piles, because they behave completely differently as your ticket grows.
The fixed pile is charged per action rather than per dollar. On an on-chain venue that means the transaction that lets the venue's contracts move your collateral in the first place, which is its own signed transaction before you have traded anything. Then the transaction that opens the position, the one that closes or claims it, and whatever it costs to get the proceeds back to wherever you keep them. Four chain interactions, none of which get cheaper because your position was small. If you are moving stablecoins on and off the chain to fund the trade, that leg belongs in the same pile.
The proportional pile scales with size: whatever the venue takes on the trade, the half spread you cross going in and coming out, and the price you move yourself if your order is large relative to the book. These are the costs people usually think about, and on small tickets they are not the problem.
I am not going to put numbers on any of this, because gas is whatever the chain charges the day you trade and venue economics differ. Measure your own. One deliberately tiny round trip on the venue you intend to use, with every cost written down, gives you a fixed figure and a proportional rate that will serve you for months.
The break-even ticket is a division
Call your total fixed cost for a full round trip F, and call your net edge per dollar traded E, after the venue's take and after the spread you expect to cross both ways. Then the ticket at which you break even is F divided by E.
Work it with placeholder numbers you should replace with your own. Suppose the four chain interactions plus the funding leg come to twelve dollars, and suppose you think you have three points of edge on a contract, of which one point goes to spread and fees on the round trip. Your net edge is two points, or 0.02 per dollar. Twelve divided by 0.02 is six hundred dollars. Below a six hundred dollar ticket, a trade with genuine two point edge loses money.
Two things about that number are worth sitting with. First, it is brutally sensitive to the edge estimate. Halve the net edge to one point and the break-even ticket doubles to twelve hundred dollars. Since most people overestimate their edge, most people's real break-even is well above the one they calculate. Second, break-even is not a target. At exactly six hundred dollars you are working for nothing. If you want fixed costs to be a rounding error rather than a partner in the trade, you want a ticket several times the break-even figure, which in this example puts a sensible minimum somewhere around two to three thousand dollars.
Where the floor meets the ceiling
Now bring in the number from the other direction. Your maximum sensible ticket is a share of the visible resting book, because above some share you are not taking the displayed price, you are creating a new one and then paying to undo it on the way out. Whatever share you pick, five percent is a defensible starting point and ten is the outer edge.
Put the two together against real readings. On the markets tab the Liquidity column across four adjacent rows shows $497.3K, $686, $648.8K and $282.0K.
- Against $497.3K, a five percent cap allows nearly twenty five thousand dollars. Your minimum of two to three thousand fits inside that with enormous room. Tradable.
- Against $282.0K, the cap is about fourteen thousand. Still comfortable.
- Against $686, a five percent cap is thirty four dollars. Your minimum viable ticket is roughly a hundred times your maximum sensible one. There is no size that works. Not a small size, not a careful size, none.
That third case is the whole point of doing this exercise. The row in question showed $796.4K of 24 hour volume, which is the kind of figure that makes a contract look alive, sitting on top of a $686 book. Volume is not depth. If you only read the volume column you would have concluded that market was busy and worth a look, and you would have been walking into a contract where the arithmetic forbids any position at all.
Measure your own numbers once, in one small deliberate trade
Here is the exercise, and it costs whatever the smallest round trip on your venue costs, which is the tuition.
Pick a contract you have no view on whatsoever, in the deepest book you can find, and take a position you can afford to lose entirely. Record the approval transaction cost separately, because you only pay it once per venue and per collateral token and it does not belong in every subsequent trade. Record the entry cost. Note the price you saw on the screen and the price you actually got, and write down the difference in probability points, because that gap is your real spread and it is usually wider than the screen suggests. Then close the position and record the exit cost and the exit slippage. Finally move the proceeds back to where you keep them and record that too.
You now have F and you have a realistic view of the spread component of E. Divide, multiply by three or four for a working minimum, and write the result on the same note where you keep your position sizing rules. Redo it if the chain gets busy or you switch venues.
The two situations that move the floor
The floor is not permanent, and there are exactly two things that shift it meaningfully.
The first is holding to resolution instead of trading out. If you never sell, you drop one chain interaction and one round of spread, which can cut both F and the proportional costs close to half. That is a real saving and it is why so much retail activity on these venues is buy and hold to settlement. It comes with a condition though: you have to genuinely mean it. Deciding at entry that you will hold to the end date, and then changing your mind in week three, is the most expensive path of all, because you paid the small ticket costs on entry and then pay the exit costs you had priced out.
The second is trading several contracts on the same venue in the same session. The approval is per venue and per collateral token, not per trade, and once collateral is sitting on the venue you are not paying to fund and withdraw for each position. A single deposit supporting five positions spreads the funding and withdrawal legs across all five, which pulls the per trade minimum down substantially. This is a real argument for concentrating your activity on one venue rather than chasing the best price across six, and it is worth checking whether the price improvement elsewhere actually exceeds the cost of getting your money there and back.
What does not move the floor is being more careful, more patient, or more right. Fixed costs are indifferent to all three, and that indifference is the entire reason to know your number before you place the ticket rather than after.