Ledgers, payouts and reconciliation
No order reaches an exchange, but the money is real: challenge fees, refunds, profit splits and payouts. Those numbers have to add up every time, including when a request is retried or a process crashes halfway through.
What tends to go wrong
- A payout retried after a timeout can be sent twice.
- Balances stored on an account drift from the transactions behind them, and nobody notices until a trader does.
- Traders and the trade press now compare payout claims with what they can check independently, so a firm's own numbers have to be right.
What I've built
Double-entry ledger
Journals are inserted as drafts and can only move to posted. Database triggers refuse a journal that doesn't balance, and posted rows can't be edited. Amounts use exact decimals.
One path for every money change
A fill, a fee, funding or a payout runs as one transaction: it claims an idempotency record and rejects a retry whose payload differs, takes per-account locks under a timeout, posts the journal and writes the events. If any step fails, none of it happened.
Updates only for committed changes
A transactional outbox makes sure the trader's terminal hears about every committed change and never about one that rolled back.
See it in the case study: Updates only for committed changes
Reconciliation
Every 15 minutes a job compares each account's stored balances with sums taken directly from the ledger, and checks journal balances and event sequences.
What I can take on
- Design a ledger for challenge fees, refunds, profit splits and payouts.
- Make payout and refund paths safe to retry.
- Add reconciliation jobs and alerts for balances that drift.
- Review an existing money path for race conditions, with tests that reproduce them.
What I haven't done yet
NEXETRO doesn't yet have a payment provider, and its payout path still needs hardening before launch.
Related work
Challenge and risk engine
Profit targets, daily-loss and drawdown floors, consistency rules, phase progression and breach detection.
Trading engine and market data
Market-data ingest, execution against real order-book depth, perpetual futures and a live trading terminal.
Back office, risk desk and dashboards
Admin panels and a risk desk for staff, trader-facing dashboards, role-based access and audit logs.
Building or fixing part of a prop firm?
Tell me which platform or market data you use, roughly how many accounts you run, and what needs to change. I'll reply with questions or a first outline of the work.