Two things happened with the trading stack this week: the finance dashboard stopped lying to itself, and we finally set a realistic goal for what we're actually trying to accomplish with it.
// Why the dashboard was lyingThe Alpaca equities bot has been live since late August. The problem is that "live" doesn't mean "honest" — the dashboard was reporting numbers that didn't match what the API was actually saying. Positions weren't reconciling, realized P&L was a best-guess calculation instead of FIFO ground truth, and the monthly goal figure ($100/month) had never been validated against actual market conditions.
This week we fixed the P&L reporter to pull realized gains the right way — FIFO from filled orders, not a rough position-level estimate — and wired the dashboard to pull fresh data from the live Alpaca account every time it loads. The numbers you see now are the numbers the API gives us.
// Setting a realistic goalWhile we were in there, we looked at what the Alpaca bot actually generates in a month. August was the first full month live. The realized P&L came in well below $100 — because August was a rough month for the swing trade setup we were running, and because $100/month on a ~$150 equity base is an aggressive target that requires either significant capital deployed or a strategy that's dialed in.
So we reset the monthly target to $20. That's not a consolation prize — it's the honest number for a swing trader at this capital level running a disciplined risk-first approach. The strategy works. The goal now matches the reality of the account.
// The 3-hour auto-refresh loopThe finance dashboard now refreshes itself every 3 hours via a cron job that pulls live numbers from the Alpaca, Kalshi, and crypto dashboards. At 9am, noon, 3pm, 6pm, 9pm, and midnight Central, the finance data file rewrites itself from the live APIs. The internal HQ dashboard reads that file, so when you open it you're seeing numbers that are at most 3 hours old, not a snapshot from whenever the last manual refresh ran.
This is the operational rhythm we want: the system keeps itself current without anyone having to remember to刷新 it. If a timer misses a beat, the cron watchdog catches it on the next cycle.
// What the lanes look like right nowThree lanes, three states:
The $20/month target is honest. We'd rather be clear about a number we can actually hit than claim a number that sets us up to either oversize risk or publish a report that reads like cope. The dashboards show the real state. The monthly report will show whether we made it. That's the deal.
Questions about the trading stack? Reach us at [email protected].