How much should I stake on a single position?

A common convention is one unit of 1% to 2% of your bankroll per position, going to about 3% only when you are unusually confident, and never more than 5% on any single outcome. With a 500 dollar bankroll that is 5 to 10 dollars a position. The point is not precision; it is that ten consecutive losses should cost you 10 to 20 percent of the bankroll rather than ending your run.

Why is a multi-leg ticket worse than separate positions?

Because the probabilities multiply. Four legs at 60% each win together 13% of the time, so such a ticket loses roughly seven times out of eight. Each leg also carries its own cost to enter and exit, and those stack. Separate positions settle independently, so one wrong leg costs you that leg rather than the whole stake.

How do I convert bookmaker odds to a probability?

Divide one by the decimal odds. Odds of 1.50 imply about 67%, 2.00 implies 50%, and 3.00 implies about 33%. The implied probabilities across a bookmaker's market add up to more than 100%, and that excess is their margin. On a prediction market the price is the probability directly: 50c means the market says 50%.

Should I increase my stake to win back a loss?

No, and this is the single most expensive habit in the list. Raising your unit after a loss means your largest stakes land during your worst run, which is exactly when your read is least reliable. The unit should move only when the bankroll moves: bigger as it grows, smaller as it shrinks.

What should I record for each position?

Date, the market and what you took, the price you paid, the size in units, and the result with the profit or loss. That is enough to answer the only question that matters at month end: which kinds of markets and which price ranges actually made you money. Most people discover they are burning money in one specific category they were sure was their strength.

GUIDE

How do I manage a bankroll on prediction markets?

September 21, 2026 · 7 MIN READ

Most people who lose money on markets are not bad at predicting. They stake three times as much after a good day, put everything on one multi-leg ticket after a bad one, and never write down what they did — so they never find out whether they are up. Five rules fix that, and none of them is about picking better.

Key takeaways

Keep a separate bankroll that could go to zero without changing your life, and never top it up from rent or bills money.
Stake a fixed unit of 1-2% of that bankroll per position, 3% at most when you are confident, so ten losses in a row costs 10-20% rather than everything.
If each leg has a 60% chance, a four-leg ticket wins 13% of the time: multiplying legs multiplies your chance of losing, and each leg adds its own cost.
Write down date, market, price, size and result for every position, then review monthly, because memory keeps the wins and quietly drops the losses.
A price is already a probability on a prediction market, so 50c means 50% — there is no odds format to convert and no bookmaker margin baked into the number.

Most people who lose money on markets are not bad at predicting. They are bad at money management — and that part is fixable with arithmetic rather than insight.

Here are five rules that decide whether you are still trading in six months. None of them will make you rich this week; that is the point.

Rule 1 — why keep a separate bankroll?

The money you trade with has to be separate from the money you live on: an amount that could go to zero without changing your life.

Decide the number once. Say $500.
Never top it up from rent, a loan, or the household budget.
When it is gone, it is gone. Do not reload to get even.

This is the rule that makes the other four possible, because every sizing rule is a percentage of something — and that something has to be fixed.

Rule 2 — how big is one unit?

A unit is a fixed percentage of your bankroll. A common convention:

Situation	Size
Ordinary position	1 unit = 1–2% of bankroll
Unusually confident	3 units maximum
Never	more than 5% on one outcome

On a $500 bankroll a unit is $5 to $10. Ten losses in a row costs $50 to $100 — painful, survivable, and you still have a bankroll to think with.

That matters more than it sounds, because long losing runs are normal rather than unlucky: at 20c a position loses four times in five by definition. Which prices produce which losing streaks is the subject of how to size a position.

The unit moves only when the bankroll moves. Bigger as it grows, smaller as it shrinks — never bigger because you are annoyed.

Rule 3 — why does stacking legs work against you?

Multi-leg tickets are attractive because the payout is large. The arithmetic is the problem: probabilities multiply.

Assume each leg is a genuine 60%:

Legs	Chance all of them land
1	60%
2	36%
3	22%
4	13%
6	5%

A four-leg ticket loses roughly seven times out of eight. And every leg you add carries its own entry and exit cost, so the costs stack alongside the falling probability.

Keep most of your volume in single positions. If you want several at once, a basket is the safer structure: each leg settles on its own, so one wrong leg costs you that leg rather than the entire stake, and you can sell a single leg without touching the others.

Rule 4 — what should you write down?

If you do not record your positions, you do not know whether you are profitable. You know which wins you remember.

Record five things per position:

Date
The market and what you took
The price you paid
Size, in units
Result, and the profit or loss

At the end of each month, look at which categories and which price ranges made money. Most people find they are losing steadily in one place they were certain was their strength — and no amount of research finds that for you, only the log does.

Give it dozens of positions before judging anything. Five trades is noise.

Rule 5 — how do you check a price against your own estimate?

Every price is a probability. Before buying, ask one question: do I think this is more likely than the price says? If not, there is no trade, however appealing the outcome.

If you are used to bookmaker odds, the conversion is one divided by the decimal odds:

Decimal odds	Implied probability	Same thing as a price
1.50	about 67%	67c
2.00	50%	50c
3.00	about 33%	33c

With a bookmaker those implied probabilities add up to more than 100%, and the excess is the margin you pay. On a prediction market the price is the probability — 50c means the market says 50% — which is the real difference explained in prediction markets vs sportsbooks. When your estimate is far from the market's, you either have a reason, or you have missed something the money already knows.

What should you check before every position?

Is this money from the bankroll, not from living expenses?
Is the size 1 to 3 units?
Am I placing this to win back a previous loss? If yes, do not place it.
Have I read the price as a probability and compared it with my own estimate?
Will I record it?

Does any of this guarantee a profit?

Prediction markets carry the risk of losing your whole stake, and no system changes that — not this one either. These rules are designed to make you lose more slowly and less often, and to make your decisions numerical rather than emotional. A position can also be closed early rather than held to zero, which selling before resolution covers.

If it stops feeling like a decision and starts feeling like a compulsion, stop and take a break. And if you have not placed a first position yet, how it works is the place to start.

Common questions

How much should I stake on a single position?: A common convention is one unit of 1% to 2% of your bankroll per position, going to about 3% only when you are unusually confident, and never more than 5% on any single outcome. With a 500 dollar bankroll that is 5 to 10 dollars a position. The point is not precision; it is that ten consecutive losses should cost you 10 to 20 percent of the bankroll rather than ending your run.
Why is a multi-leg ticket worse than separate positions?: Because the probabilities multiply. Four legs at 60% each win together 13% of the time, so such a ticket loses roughly seven times out of eight. Each leg also carries its own cost to enter and exit, and those stack. Separate positions settle independently, so one wrong leg costs you that leg rather than the whole stake.
How do I convert bookmaker odds to a probability?: Divide one by the decimal odds. Odds of 1.50 imply about 67%, 2.00 implies 50%, and 3.00 implies about 33%. The implied probabilities across a bookmaker's market add up to more than 100%, and that excess is their margin. On a prediction market the price is the probability directly: 50c means the market says 50%.
Should I increase my stake to win back a loss?: No, and this is the single most expensive habit in the list. Raising your unit after a loss means your largest stakes land during your worst run, which is exactly when your read is least reliable. The unit should move only when the bankroll moves: bigger as it grows, smaller as it shrinks.
What should I record for each position?: Date, the market and what you took, the price you paid, the size in units, and the result with the profit or loss. That is enough to answer the only question that matters at month end: which kinds of markets and which price ranges actually made you money. Most people discover they are burning money in one specific category they were sure was their strength.