Orca Odds

One ledger for every bet.

You bet on four apps.

What you can do

  • Build an agent

    Start from a house agent or a blank page: choose how it picks, its leagues, price range and picks a day. It runs every morning on real games and is graded with no money down. Three on the free plan.

  • Climb the leaderboard

    Agents are ranked on the price they beat, not on how often they win, with the sample size beside every rank. Your agents' combined record is your own place among the people who build them.

  • Log what you bet

    Log a slip against the pick it followed, sync Kalshi and Polymarket, or import a statement from any sportsbook. One ledger that is never edited, only added to.

  • Join the community

    Compare agents, comment on a record, fork one that works and make it your own, and ask the forum what you are missing.

Meet the house agents

Three personalities, one record. Each picks within its own limits and is judged the same way.

Orca-Alpha-1
The value hunter
“I don't predict. I price.”

Takes the market's own view of a game from every book it can see, with the bookmaker's cut removed, and backs a side only where one book pays more than that view says it is worth.

  • Shops every book
  • Price over opinion
  • Patient
Up to 15 picks a day · prices −250 to +1000 · 9:00 ET
Orca-Beta-1
The favourites backer
“I like winning. I just won't overpay for it.”

Backs sides the market gives at least a 60% chance, at the best price on offer, and pays no more than the usual bookmaker's cut over fair. Its record is still judged by price, not by how often it wins.

  • Plays the likely side
  • Fair price only
  • Steady
Up to 10 picks a day · prices −300 to +150 · 9:00 ET
Orca-Delta-1
The crowd follower
“Follow the money, then check the price.”

Ranks the day's games by how much Kalshi and Polymarket traders have put on them and, for each of the busiest, backs the side with the better price against the market's own view.

  • Most-traded games
  • One pick per game
  • Price-checked
Up to 30 picks a day · prices −500 to +500 · 9:00 ET
Orca-Gamma-1
The researcher · coming soon
“Give me the slate and the headlines. I'll tell you why.”

Reads the day's slate and the news around it, picks only the games it is sure about, and says why in plain words. It is the only agent that uses a language model.

  • Reads the news
  • Picky
  • Explains itself
Up to 3 picks a day · 9:30 ET

A record you can audit

Every agent, and every slip you log, goes round the same four steps, and each step is kept.

  1. 1Lock

    A pick is locked at a real price before the game starts, and cannot be changed afterwards.

  2. 2Grade

    After the final: won or lost, and whether the price beat the closing line.

  3. 3Track

    Results are kept for each version of an agent, so a change earns its own record.

  4. 4Compare

    A slip you log credits the pick it followed, so you and the agents share one ledger.

About Orca Odds

We are building the honest scoreboard for betting: agents you build, graded on the price they beat, and one record of what you bet beside them.

The team

  • Anthony TrevinoFounder
  • We never place your bet

    You place it at the venue and log it. We only keep the record, so nothing we show can nudge a stake into the market.

  • No scraping, ever

    Accounts come in through official APIs and statements you import. We do not break a platform's terms to get your data.

  • Win rate is not the score

    We rank on the price an agent beat and show the sample size beside it, so a lucky week cannot pass for skill.

Get on the list

Orca Odds is invite only while we open places in small groups. Leave an email and we will write when yours is ready.