Documentation

πŸ€– Autobond

Bonds at 97% and up do not sit around waiting for you. They appear when news lands, and the good ones are taken within minutes. Overnight, most of them come and go while you are asleep.

Autobond watches for you. You set a budget and the rules you are comfortable with, and it buys the bonds that fit, sells them if they turn against you, and stops when it has lost more than you said it could.

Open it from Bonds, or go straight to Autobond.

What you set

Six things decide everything Autobond does.

Budget. The most it will have tied up at once. Your money stays in your wallet until it actually buys something.

Positions at once. How many bonds it can hold. Your budget is split evenly between them, so a $300 budget across 3 positions puts $100 into each.

Each position needs at least $5. Polymarket will not take an order smaller than 5 shares, so a budget that leaves less than that per position cannot buy anything, and Autobond will not let you save it. Give it a bigger budget or fewer positions.

Lowest odds to buy. It will not buy anything the market rates below this. The floor is 97% and you cannot go under it.

Lowest return to accept. Expressed per year, so short bonds show large numbers. See why that can mislead.

Settles within. Only bonds that finish inside this many hours are eligible.

Sell if it drops. Your stop loss, in cents. See what it can and cannot do.

You can also limit which subjects it buys, which risk ratings it accepts, and whether it may hold two positions on the same event.

How Autobond picks a bond

Every so often it looks at the bonds currently on offer and drops any that fail one of your rules. What is left has to also be buyable: there has to be someone selling, at a price that still meets your odds and return thresholds, in enough size to fill your per-position amount.

If nothing survives, it buys nothing and waits. That is the normal case, not a fault. The panel titled Why Autobond has not bought on the Autobond page shows you exactly which rule ruled out the most bonds on its last look, and which bond came closest.

Loosening one rule at a time is the fastest way to widen the pool. Dropping your odds threshold from 99% to 97% usually has the biggest effect, and it is also the one that raises your risk the most.

Why APR can mislead

A bond that pays 1% and settles tomorrow is roughly 365% a year. The same 1% over a month is about 12%. Nothing about the trade changed, only the clock.

That is why a very high return threshold does not filter for better bonds. It filters for bonds that settle soonest, and for the ones other people are not buying. If a bond is offering far more than its neighbours at the same odds, the usual reason is that the market is less sure than the price suggests, or that there is very little of it to buy.

Your odds threshold and your risk ratings do more to control your actual risk than the return threshold does.

Fixed or reinvested profits

Keep them. Autobond keeps aiming at the budget you set. Profits build up in your wallet and are not put back to work. If you lose, it will use other available funds in your wallet to get back to your budget.

Reinvest them. Autobond adds your profits to its target and subtracts your losses. Positions grow as you win and shrink as you lose.

A worked example. You set a $1,000 budget across 4 positions, so $250 each.

  • On Keep them, after making $200 you still have a $1,000 target and $250 positions. The $200 sits in your wallet.
  • On Reinvest them, that $200 raises the target to $1,200, so positions become $300. After a $150 loss the target falls to $1,050 and positions become about $262.

Reinvesting compounds in both directions. It is the more aggressive choice.

What the stop loss can and cannot do

If you set 3 cents, Autobond tries to sell when the best price anyone is offering falls 3 cents below what you paid.

What it does well: it caps the slow bleed, where an outcome drifts from 98Β’ to 95Β’ to 90Β’ as the market changes its mind over hours.

What it cannot do:

  • It cannot beat a settlement. Most bonds that go wrong do not drift. They resolve. A 98Β’ outcome that loses goes to zero at settlement, with no path down for a stop loss to catch.
  • It needs a buyer. Selling requires someone on the other side. In a thin market at a bad moment, there may be nobody at your price, and the sale can take several tries or fail.
  • It sells at the going price, not your trigger. The price it gets can be worse than the one that set it off.

Leaving it empty turns it off completely. Autobond will then hold every position until it settles, or until you sell it yourself from the Holding now tab.

Why losses can pass your limit

Stop buying after losing is a circuit breaker on new buys. When your realised losses reach it, Autobond stops opening positions. It does not sell what you already hold.

So your total losses can end up larger than the number you set. Positions that were already open when the limit hit are still exposed, and they still settle on their own terms. If you want out of those too, sell them yourself from the Holding now tab.

The same is true of Stop Autobond: it stops new buying, and leaves your open positions running with their stop losses and settlement handling intact.

Following what it did

Holding now lists what Autobond currently owns for you, what you paid, and what you make if each one wins. A row highlighted in amber or red is trading below what you paid.

Finished lists what has settled or been closed, with what you made or lost after fees.

What Autobond did is the running log: every buy, sale, payout, and stop.

Each wallet runs its own Autobond, with its own budget, its own rules, and its own results. Switching wallets at the top of the page switches all of it.

The honest part

These are still bets. A 98% outcome loses about one time in fifty, and when it loses you lose the whole position, not 2% of it. Run enough of them and you will meet that one.

Autobond does not change those odds. It only saves you from watching a screen. Set a budget you would be comfortable losing, keep the loss limit low while you are learning, and give it small positions until you have seen it run through a losing bond.