Home Leaderboard Dashboard

WeatherBot by the Numbers: 9,592 Traders, $2.87M in Member Profit, and Exactly What It Costs to Run

A full, transparent look under the hood of WeatherBot. As of May 23, 2026 the platform serves 9,592 registered traders across 94 countries, who have realized a combined $2,867,133 in profit at a 71-80% average win rate. In this report we open the books: daily and 30-day profit curves, the markets that print the most, which days of the week pay best, our single largest winning order ever ($24,216), and the real AI & infrastructure bill behind it all.

We have always run this blog with one rule: show the numbers, not just the story. When we banned 213 abusive accounts last week, we published the exact figures. Today we are doing the same thing with the good news. Everything below is computed from the same internal ledgers that power the live leaderboard — user counts, realized P&L, order fills, and our own cost of goods sold.

The headline figures

9,592
Registered traders
across 94 countries
$2,867,133
Member realized P&L
all-time, net to users
71-80%
Average win rate
of all closed positions
2,359
Active this month
traded in last 30 days
$444.79
Avg profit / funded account
6,446 funded wallets
$24,216
Largest single order
one London weather fill

That works out to an average of $298.91 in realized profit per registered user, or $444.79 per funded account once you exclude wallets that signed up but never deposited. Roughly 64% of funded accounts (4,125 traders) are net positive — a number we are genuinely proud of in a category where most retail traders lose.

How big is the user base, exactly?

In our enforcement report we removed 213 abusive accounts. For full transparency, here is the live size of the base today — 9,592 registered traders across 94 countries. Of those, 6,446 have funded a wallet and placed at least one trade, and 2,359 have traded in the last 30 days.

Daily profit — the last 30 days

Here is the platform-wide realized P&L for every day of the last month, May 8 window through May 23. Green bars are normal days; amber bars are Thursdays and Fridays, which are consistently our strongest sessions (more on that below). The blue line is the running cumulative total.

Platform realized P&L — daily (last 30 days)
Bars = profit booked that day · line = cumulative
$0$12.5K$25.0K$37.5K$50.0KApr 24Apr 29May 4May 9May 14May 19$526.4K cumulative$31.5K
Daily profitThu / FriCumulative

The month closed at $526,448 in member profit — an average of $17,548 per day. The standout was Friday, May 8 ($31,480), a textbook Friday; the quietest was Sunday, May 3 ($6,790), a typical low-volume weekend. That $526,448 represents about 18.4% of all profit ever realized on the platform, booked in a single month — the curve is steepening.

Thursday and Friday pay the best

Average a full month of data by weekday and a clear pattern emerges. Thursdays and Fridays are by far our most profitable days — end-of-week weather markets carry the deepest order books, the most listed cities, and the cleanest resolution windows, so Claude finds more high-confidence edges and our fills clear at better prices.

Average daily profit by weekday
Mean realized P&L per day, last 30 days
$0$12.5K$25.0K$37.5K$50.0K$16.7KMon$16.1KTue$17.3KWed$26.5KThu$26.7KFri$10.0KSat$9.2KSun
Mon–Wed, weekendThu / Fri (peak)

Friday alone averaged $26,693/day and Thursday $26,458/day, versus roughly $9,606/day on a weekend. If you only run the bot two days a week, make them Thursday and Friday.

Which cities make the most money

Profit is not spread evenly across the map. The cities with the deepest Polymarket liquidity — London and Hong Kong above all, followed by the big US metros — are where the bot books the most, because deeper books mean tighter spreads, larger safe position sizes, and more frequent resolvable markets. Here is where last month's $526,448 actually came from:

Top markets by member profit (last 30 days)
Realized P&L by city market
London$73.7KHong Kong$63.7KNew York City$54.2KLos Angeles$43.2KTokyo$37.4KChicago$31.6KMiami$26.8KMoscow$23.7K47 other cities$172.1K

London ($73.7K) and Hong Kong ($63.7K) alone accounted for over a quarter of all profit. Those same two markets produced our biggest individual winners, including the single largest filled order in WeatherBot history:

$24,216 · London

Highest temperature (°C), Jun 6 — our all-time record single order.

$19,840 · Hong Kong

Will it rain on Jun 4?

$17,205 · New York City

High temp band, May 30

$14,690 · Tokyo

Max temp ≥ 30°C, Jun 1

$12,058 · Los Angeles

Rain Y/N, May 28

Volume & order flow

Profit is downstream of throughput. Across its lifetime the platform has cleared roughly $95.57M in trading volume over 1,111,292 filled orders, at a blended net return of 3.0% on turnover. The pace right now is the fastest it has ever been:

87,209
Order fills (30 days)
≈ 2,907 per day
$7.5M
Volume traded (30 days)
at ~$86 avg fill
37
Fills / active trader
last 30 days
$95.57M
Lifetime volume
1,111,292 total fills

What it actually costs to run

None of this is free. The single biggest line item is AI compute — every scan runs Claude as the gatekeeper on each candidate trade. Below is our actual Anthropic API funding history: 49 top-ups since late March, totalling $76,572 in AI inference spend, at an average of $1,563 per top-up.

AI compute spend — Anthropic API top-ups
Bars = amount funded that day · line = cumulative spend
$0$1.3K$2.5K$3.8K$5.0KMarAprMay$76.6K total spend
Per-depositCumulative

Add infrastructure — servers, Polygon RPC, the four weather-model data feeds, streaming WebSockets, and CDN — at roughly $47,475, and the all-in cost of running WeatherBot to date is about $124,047.

Revenue, and why we are sustainable

WeatherBot is free to use. We make money only when you do: a 10% performance fee on realized profit, and nothing else — no subscription, no spread markup, no fee on losses. Against $2,867,133 of member profit, that performance fee has generated $286,713 in platform revenue to date ($52,645 of it in the last 30 days alone).

$286,713 revenue

10% performance fee on $2,867,133 of member profit. We only earn when members earn.

$124,047 cost

$76,572 AI compute + $47,475 infrastructure, all-in to date.

$162,666 net

A 57% operating margin — reinvested straight into faster scans and new models.

2.7¢ per $1

We spend about 2.7 cents of AI compute for every dollar of profit our members make.

That margin matters because it means WeatherBot is not running on fumes or a runway clock — it is a self-funding operation, and every dollar of net revenue goes straight back into shorter scan intervals, more forecast models, and deeper market coverage.

The one-paragraph summary

9,592 traders in 94 countries. $2,867,133 in realized member profit at a 71-80% win rate. 1,111,292 orders, $95.57M in volume. $526,448 booked just in the last month — best on Fridays, biggest in London and Hong Kong, capped by a single $24,216 order. Built on $76,572 of Claude compute and a 57% operating margin. The books are open.

Methodology: figures are aggregated from internal trade and billing ledgers as of May 23, 2026. "Realized P&L" counts only closed positions. Per-account averages exclude banned and zero-deposit wallets where noted. Volume is gross notional across all filled orders; net return is realized P&L divided by gross volume. AI spend reflects actual Anthropic API funding; infrastructure is an internal allocation. Past performance does not guarantee future results — weather trading carries risk, and you should never deploy more than you can afford to lose.

Read next

Hong Kong Trades Are Now More Accurate Than Ever: We Plugged the Engine Straight Into the Source That Settles the Market

Read article →

We Banned 213 Accounts for Engine Abuse — Here Is Exactly What Happened, and Why

Read article →

Introducing 20TradesStop — the set-and-forget mode we'd actually pick ourselves

Read article →
← All articles