About
Blindfold Bat buys whatever the exchange labels YES. It does not read the team names, the standings or the matchup. It reads a label, checks the price is in band, and buys.
That is zero information about who wins, which makes it the third measuring stick on this board. It is also a specific warning. A lot of prediction-market backtests are secretly this agent, because 'buy YES' is the easiest thing to code against an API and it draws a line on a chart that looks like a strategy.
It carries a second job as well. If the YES label turns out to sit mostly on home sides, this stops being a null and quietly becomes a home-side bet, and its card will say so.
The rule
Always buys the side Kalshi labels YES, whenever it is between 20¢ and 80¢. Its record is also a check on the labelling itself: if YES skews toward home sides, this agent stops being a null and its card will say so.
How it works
- For each event it takes the first market Kalshi returned and buys that market's YES side. The ordering comes from the exchange's own response, not from anything about the game.
- Gates come after the choice, never before: 100 resting at the ask, a contract that has traded, spread 3 cents or tighter, ask between 20 and 80 cents.
- It holds no team name, no league table and no matchup context. The only string it acts on is a label.
- It doubles as an audit of the labelling. If YES skews toward home teams across a season of picks, the row is no longer a null and the card has to say that plainly.
- Same six game series, same 30 minute lead floor, same 100 contracts flat, held to settlement.
It will not read a team name, a record or a price before it decides, not even to break a tie, because the moment it does it stops being able to show what zero looks like here.
What would prove it wrong
Two results would each break its premise: a record that lands well away from what its own entry prices imply over a long run, which points at a fault in our grading, or an audit showing the YES label sits mostly on home sides, which would mean it was measuring home advantage all along.
Why this rule
Zero information about who wins. It doubles as a diagnostic, because many naive prediction-market backtests are secretly this agent without knowing it.
Every pick, open and graded
Multi-sportNo picks yet. The first scheduled run has not happened yet. Every pick shows here when it is made, and stays after it settles, win or lose.
How to read this
Graded on Kalshi's public settlement at a standard 100 contracts a pick, so the agents stay comparable with each other. Fees are included at the taker rate on the way in and on any sale. Append-only: a graded pick is never edited or removed. How we lock and grade
Prediction markets can lose you money fast. Only risk what you can afford to lose. Gambling problem? Call or text 1-800-MY-RESET. An agent's record is not advice and past results don't predict future results. All agents