Guaranteed return: when it exists, and why it almost never does
Your data
These are meant to be the best prices you can actually get for each outcome, and in practice that means different houses. Prices from the same house will not work, and the results below explain why in one line.
Results
| The implied chances added together, in percent | — |
| Guaranteed return, in percent | — |
| The same in money | — |
| Odds the other side needs to pass, for a two way market | — |
How the money has to be split
| Outcome | Decimal odds | Money on it | What comes back if it wins |
|---|
The last column is the same number on every row, and that is the whole trick: the money is split in proportion to one divided by each price, so the payout is identical whichever way the event ends. Split it any other way and the returns stop matching, which turns a covered position back into a bet on one side.
Now the part that matters more than the arithmetic. The condition for any of this is that the implied chances add up to less than a hundred, and inside a single house they never do, because adding up to more than a hundred is what the margin is. A pair priced at minus one hundred and ten on both sides comes to 104.76.
Where it starts, by price
| One side is priced at | The other side has to pass | Which is exactly a hundred, in percent |
|---|
So the requirement is not a clever bet, it is two houses disagreeing by more than each of their margins. With two point one on one side, the other side has to beat one point nine zero nine, and any price under that is an ordinary bet wearing a better name.
It also explains why the returns are small when they exist at all. The prices this page starts from are a wide disagreement by real standards and they pay under four per cent, which is what is left after two houses have both taken their cut and still ended up on opposite sides of the truth.
What the arithmetic cannot see is the part that decides whether any of it survives contact. Both prices have to still be there when the second bet is placed, both houses have to accept the size, and accounts that keep doing this tend to stop being allowed to. The number above is a ceiling, not a plan.
When does covering every outcome actually pay?
Only when the implied chances of every outcome add up to less than a hundred. That is the whole condition, and everything else on the page follows from it.
Inside a single house they never do, because adding up to more than a hundred is exactly what the margin is. A pair priced at minus one hundred and ten on both sides comes to 104.76, so the thing being looked for is two houses disagreeing by more than each of their cuts.
How exact does the split have to be?
Exact. The money goes in proportion to one divided by each price, which is what makes the payout identical whichever way the event ends.
Split it any other way and the returns stop matching, and a covered position quietly becomes a bet on one side. The error does not shrink the profit, it changes what the operation is.
| One side is priced at | The other side has to pass |
|---|---|
| 1,50 | 3,000 |
| 1,80 | 2,250 |
| 2,00 | 2,000 |
| 2,10 | 1,909 |
| 2,50 | 1,667 |
| 3,00 | 1,500 |
Why is the return so small when it does exist?
Because it is whatever is left after two houses have both taken a cut and still ended up on opposite sides of the truth. The prices this page starts from are a wide disagreement by real standards and they pay under four per cent.
That is also why it is worth checking rather than assuming: a gap that looks generous on screen is often a rounding away from paying nothing.
So can I just do this repeatedly?
The arithmetic says nothing about that, and the parts it cannot see are the ones that decide. Both prices have to still be there when the second bet goes on, both houses have to accept the size, and accounts that keep doing this tend to stop being allowed to.
What the page gives you is a ceiling on one specific pair of prices, not a plan and not a recommendation.
Reactions
0
0 Comments
Be the first to comment