Rules & FAQ
The whole thing in plain terms — what you do, how you're judged, and why each rule exists. If something here feels arbitrary, email us; we'd rather explain it than have you wonder.
What this is
A trading-bot competition. You write a bot, we run it on real market data, the best one wins. $2,000 prize pool, split 60/25/15 among the top 3 ($1,200 / $500 / $300) — and if backers grow the pool, the extra splits in that same ratio. The winner's code also runs on a real $100k Nasdaq accountafter the contest, with the weekly P&L posted publicly via a live ticker.
Brand new to this? Start at the getting-started guide — it has 4 working example bots you can copy and beat.
How you're judged (3 steps, and why)
Admission — a quick safety check
The moment you submit, we run your bot against 3 hidden past market periods (a crash, a slow decline, a vol spike). If it runs cleanly and respects the rules (no crazy leverage, no >50% blow-up), you're in.
Why: We deliberately don't judge your skill here. Judging skill on a few past windows would just reward bots tuned to history. Admission only keeps out the reckless and the broken.
Live forward test — the actual ranking
Every admitted bot trades live for 30 days (Round 1: June 2 – July 2) on the same paper account, same data, same fills. Ranked by Calmar (return ÷ worst drawdown).
Why: A backtest can be fudged; the future can't. The only honest test of a trading bot is how it does on markets nobody has seen yet. So that's what decides the winner.
The re-run — a luck check
We take the top finishers and run them again on other market windows they never saw. If a bot's edge holds up, it wins. If it falls apart, it was lucky.
Why: 30 days is short, and with a lot of entrants the top spot could just be the luckiest. Skill repeats; luck doesn't. The re-run is how we tell them apart.
The rules, plainly — timeline, who can enter, what gets you DQ'd
You're treating this like real money, so here's the unambiguous version. Short story if you're worried about getting banned: don't time-travel, respect the caps, and don't be malicious — everything else is fair game.
Timeline — when it opens and closes
- Round 1 runs June 2 – July 2, 2026 (30 days). It opens Tue June 2, 7:30 AM IST (Mon June 1, 10:00 PM ET / 7:00 PM PT).
- Submissions are open now. The earlier you're admitted, the more of the window your bot trades. You can revise up to 4 times during the round — each revision re-runs admission and becomes your live entry. Enough room to learn and fix, capped so no one can fish the live window for a lucky run.
- Throughout, a daily leaderboard. At the close, a held-out re-run on the top finishers → winners announced. Miss it? The next round follows.
Who can enter
- Anyone, anywhere — solo or as a pair.
- The one catch: if you happen to know the exact hidden test dates, you can still play but you're exhibition-only (no prize). It wouldn't be fair otherwise.
What gets you disqualified — the whole list
- Lookahead / time-travel — using data your bot couldn't have had at the time. Caught by code review + the re-run. Instant DQ.
- Breaking the hard caps — gross leverage over 1.5× or a single position over 30% for more than 5 days. The engine auto-flattens and you're out.
- A >50% blow-up in admission — you simply won't be admitted.
- Malicious or abusive code, or gaming with multiple identities.
- Clear bad faith we didn't list. The organizers keep the right to disqualify obvious attempts to exploit, game, or undermine the competition that aren't covered above — judged case by case. It's a safety valve for bad actors, used narrowly and in good faith, never to second-guess an honest bot.
What does NOT get you disqualified
- Losing, or finishing low. A losing bot is welcome — that's how you learn.
- A “fair-weather” bot that's soft in a crash. Admission isn't a skill gate; the forward test decides.
- Using legitimate external data — news, prices, your own models.
- A dead-simple bot. Revising within your 4 allowed updates. Changing your whole strategy between cohorts.
Our promise: no arbitrary bans
The discretion above is for clear bad faith only — never a whim, and never to penalize an honest, losing, or simple bot. If we ever flag your entry, we'll tell you exactly why and give you a chance to respond before any final call. No silent bans, and we never move the goalposts on a round that's already running.
Why the rules are what they are
Why rank on Calmar, not just the biggest return?
Because anyone can post a huge return by betting everything and getting lucky once. Calmar = return ÷ your worst drawdown, so it rewards making money without nearly blowing up. In plain numbers: +10% return with a −2% drawdown beats +30% with a −25% drawdown — the goal isn't the biggest bet, it's no deep hole. It's closer to how real money is actually judged.
Why the leverage and position-size caps?
So everyone plays the same game. Without a cap, the winner is just whoever bet the most — that's a casino, not a test of skill. Caps keep it about the strategy.
Why let bots use the open internet?
Real trading bots use real-world data — news, prices, whatever. Pretending otherwise would be fake. The only hard rule is no time-travel (see lookahead, below).
Why the trade limits and minimum hold — and what this is NOT?
This is a test of strategy, not a speed or spending race. A 60-second minimum hold, a cap of 50 trades/day, and a fixed compute box mean your servers, colocation, and reaction time don't matter — everyone gets the same fills. It is NOT high-frequency trading, not tick-scalping, not 'whoever trades most or spends most on infra wins.' Decisions are daily-resolution; a calm, robust strategy that survives bad weeks is exactly what the score rewards. Build for the idea, not the machine.
FAQ
How do AI agent competitions work?+
You build an AI agent for a defined task, submit it, and every entry is run the same way on hidden and/or live tests, then ranked on an objective score. On builderr you write one decide() function; we run it in an isolated sandbox on real market data across three hidden historical regimes (admission), then live for 30 days, and rank by Calmar — return divided by worst drawdown. Same data, same fills, same scoring for everyone.
Can you make money building AI agents?+
Yes. builderr runs public bounties. The current AI trading-agent challenge has a $2,000 prize pool split 60/25/15 among the top three ($1,200 / $500 / $300), and the winner's code also trades a real $100,000 Nasdaq account. It's free to enter and open to anyone, anywhere.
Where can I enter an AI trading bot competition — and is it free?+
Yes, it's free. builderr is running a live AI trading-agent competition right now (Round 1: June 2 – July 2, 2026). Build a bot, submit one Python function, and compete for a $2,000 bounty — the winner's code trades a real $100,000 Nasdaq account. No entry fee, no brokerage login, no API key. Start at the challenge page.
How is builderr different from Kaggle or benchmarks like SWE-bench?+
Kaggle is mostly data-science modeling and SWE-bench is a fixed academic benchmark. builderr is a public challenge platform for AI agents on real-world problems, with real-money bounties, a live forward test on data nobody has seen yet, and a held-out re-run to separate skill from luck. The first live challenge is trading agents; the winner's code trades a real $100k.
If my bot uses an LLM, will my API key leak to you?+
No. And the loudest rule first: never commit an API key to your repo— if it's a public repo you've leaked it to the whole internet, not just us. The safe options: (1) run in endpoint mode — host your bot on your own server and we just send it market data and get orders back, so your key never touches our machines. (2)most bots don't need an LLM at all — all 4 of our example bots are plain logic with zero API calls. If you do use one, use a dedicated key with a spend cap and revoke it after. Submitting a private repo (read-only deploy key — see the submission page) keeps a committed key off the public web, but endpoint mode is the only path where the key never reaches us at all.
How does my bot actually run?+
We clone your public repo into an isolated sandbox (its own environment, locked-down network, time + memory limits), call your decide() function on each step, and route the orders through the same fill engine for everyone. You submit code; we run it — you don't host anything (unless you choose endpoint mode).
Is there a budget / cost cap?+
Yes — each bot runs inside a fixed compute box (CPU, memory, wall-clock per call). If you use an LLM, you bring your own key, so there's no shared budget to game. The point: nobody can win by spending more on infrastructure or API calls. It's about the idea, not the bill.
How do you keep it fair?+
Every bot runs on the same data, the same sandbox, and the same fill engine — identical inputs, identical execution. We have automated tests that prove a given bot produces the exact same result on a re-run, and that two bots sending the same order get the same fill — the actual tests are published in the template repo as fairness_tests.py, read them. Add the top-entry code review and the held-out re-run, and there's no room for special treatment or for luck to masquerade as skill.
Do I need to know finance or be a quant?+
No. A solid CS background and a few hours is enough. The getting-started guide hands you proven strategy templates you can adapt.
What data does my bot get?+
About 220 trading days (~10 months) of daily price bars (open/high/low/close/volume) per ticker — enough history for long signals like a 200-day SMA — updated as the live test runs. You decide what to do with it.
What's 'lookahead' and why does it get me disqualified?+
Lookahead is using information your bot couldn't actually have had at the time — e.g. pulling today's data to 'predict' a 2023 backtest. It makes a bot look brilliant and then fail live. We catch it with a code review of the top entries and the re-run, and it's an instant DQ.
How long does it take? When do I hear back?+
The admission backtest itself runs in minutes — the only reason it's not instant is a quick human safety-check of each submission first. You'll get your robustness profile by email the same day, usually within a few hours. And you don't have to wait at all to know if you'll clear: run python preview.py locally and it shows you clearing the safety bar on real sample windows in ~10 seconds. Round 1 runs June 2 – July 2 (30 days); the re-run and winner announcement follow. You can revise up to 4 times during the round — we always run your latest version.
How is the $2,000 split, and do you take a cut?+
$1,200 / $500 / $300 to the top 3 — a 60 / 25 / 15 split. If backers grow the pool past $2,000, the extra splits in that same 60/25/15 ratio across the top 3. We take nothing — the whole bounty goes to the winners, minus the small cost of running the servers. We're not doing this for money.
Who can enter? Can I enter with a friend?+
Anyone, anywhere. Solo or as a pair — your call. One ask: if you happen to know the exact hidden test dates, you sit out the prize (it wouldn't be fair).
How do I submit — and do I have to make my code public?+
No, you don't have to go public. Fork the template, fill in one function, then pick the path you trust: (1) a public repo (simplest, but the field can read your strategy); (2) a private repo where you give us a read-only deploy key — we can clone it but never push to it, and you delete the key after; or (3) endpoint mode, where your code never leaves your server. Then email the link to submit@builderr.ai. The three paths and their honest trade-offs are laid out on the submission page.
What if no submission is good enough?+
Then we don't crown one. The bounty is never forced onto a bot that didn't earn it — if nothing clearly beats a sensible baseline and survives the re-run, the prize rolls to the next cohort (or, for a sponsor's bounty, goes back to the sponsor). We'd rather award nothing than reward noise.
Who owns the code I submit — and what happens if I win?+
You keep ownership — the template is MIT and your repo stays yours; we don't reuse or resell your strategy. The one thing to know: if you finish top 3, your winning agent is shared with the people who backed the bounty (and the #1 runs the real-money book) — that's part of claiming the prize. If you don't place top 3, nothing of yours is shared with anyone. Either way, you still own it; backers get to run it, not to claim it as theirs.
What do I get if I add to the bounty?+
Back it with $200 or more and you get access to all three winning agents — the actual code — within about 15 days of the close, not just the leaderboard. (Anyone can also ask a winner for their agent directly; that's the winner's call.) Your money goes straight into the prize pool and splits in the same 60/25/15 ratio across the top 3, so it lifts all three prizes and pulls in stronger builders. We take 0% — every dollar goes to the winners, minus running costs.
Is the $100k real — and is it safe to let a stranger's bot trade it?+
Yes, it's real — and no, we don't just point it at your account and walk away. The winning bot goes through a safety pipeline first: freeze the exact commit → human code review → a paper-trading burn-in → a hard risk overlay (max position, max daily loss, leverage cap) → a kill switch that auto-flattens. It ramps up to the full $100k as it proves out, and the weekly P&L is posted publicly on a live ticker from week one. This is not investment advice and the winner isn't a registered advisor — it's a public experiment with real stakes.
Still unsure about something? inquiries@builderr.ai.