Follow-up to 77d8c2e from its adversarial review:
- Stop-after-all no longer shuts down before an open while histories are
still loading (a regression in 77d8c2e): a pending account counts as in
play everywhere but the weekday 2:55-5:00 PM CT stretch after the close.
- Tradovate balance holds lift only when a Cash History books a fund
transaction the account didn't have when the hold began; one that never
does is released after 3 clean reports over 15 minutes, with a warning.
The baseline comes from the balance when the report was requested, so a
payout landing mid-report is held for the next one. Held accounts
refetch every 5 minutes, not every sync.
- The Cash History baseline is kept in account_meta, so a restart onto a
failing report server still catches a payout from the DB cache.
- A wipe holds the account until a Cash History rebuilds its fund
transactions (the Fills report and the cache can't).
- A report still queued after its polls, or answered without data, is a
failure, not an empty history.
- TopstepX: a newly fetched account (new, wiped, failed) waits for a sync
that reconciles its funds, and the fetch starts that sync right away.
- Accounts a fetch is already working through aren't fetched again by the
targeted pass; the TopstepX REST fill search spans ±15 minutes.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Tradovate rolls each account's cash record to the new trade date on its own
schedule (18 minutes after the 17:00 CT open on 2026-09-24, or not until the
account trades), and until then the latest record still carries the ended
day's realizedPnL. Plans sized legs off it, one-and-done accounts read as
already traded, the reporter scored every account partial/loss, and an entry
across the record roll took a whole day's P&L as its "commission", parking
the take-profit out of reach.
- Cash records keep their trade date; realized P&L from an earlier trade
date reads as 0 (judged when read, so a record stored before the roll
counts as ended after it). Archived or older records never overwrite a
newer one, whichever order the roll pair arrives in.
- The entry-commission delta falls back to $2.50/contract when it exceeds
any real fee ($10/contract).
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
- Both clients track which accounts this process has loaded history for
(historyReady). An account restarted, rebuilt, wiped or newly listed is
ineligible until its fetch lands, instead of trading as a fresh stage-1
first day; a failed fetch with no cache no longer reads as empty history.
Waiting accounts get a targeted fetch on the next sync.
- Tradovate holds an account whose start-of-day balance moves without a fill
(a payout between hourly Cash History runs) until a Cash History books it,
so a payout-pending account isn't re-traded on its ended cycle.
- Wipe routes forget history through the client and count accounts whose
history didn't come back as failures.
- TopstepX sendOrder picks up a fill that arrived before the place response,
and its REST backstop uses the documented Trade/search and Order/search.
- POINT_VALUES: SI 5000 (was 50), M2K 5 (was 10).
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Why the autotrader sometimes broke after a weekend until `pm2 restart`:
`next start` only logs uncaught errors, so nothing ever crashed, and a few
paths could leave a client dead for the life of the process.
Tradovate: a closed socket scheduled exactly one re-login, with no catch —
one failure during weekend maintenance (or a rate-limit penalty returned as
a 200 without a token) and the client never came back. connectAndAuth is now
a single-flight loop that never throws: jittered backoff capped at 5 min,
penalty tickets waited out, captcha demands backed off. A 15s watchdog
abandons silent, never-authorized or stuck-closing sockets and stale syncs;
renewal falls back to a fresh login; orders fail fast (OrderNotPlacedError)
on a dead socket and time out instead of hanging; a bad sync answer no
longer wipes accounts and positions.
TopstepX: hub start/stop/invoke and every REST call are bounded; a wedged
hub is discarded and rebuilt; a stuck sync is superseded after 3 min;
re-login no longer tears down the sync loop; a watchdog re-logs when syncs
go stale; a no-fill timeout cancels the order.
Scheduler: one unready client used to hold every firm forever. It now holds
the tick for a 10-minute grace, then trades the other firms without it.
Per-firm try/catch in batch collection, one batch at a time, accounts marked
in flight, legs un-counted only when the entry never reached the broker, and
the scheduler starts before the first batch fires.
Boot and supervision: instrumentation.ts boots the process at server start
(it used to wait for the first dashboard request); a supervisor rebuilds a
client that has been unhealthy for 20 min (not a TopstepX one holding a
position); broker HTTP calls default to a 60s timeout; GET /api/health shows
each client's liveness and the scheduler's last decision. Shared state the
eager boot would otherwise split across bundles lives on globalThis. Logs no
longer carry access tokens or whole AxiosErrors (which include the encoded
Tradovate password).
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
TopstepX never reports the Maximum Loss Limit, but it's deterministic:
floor = min(0, eodBalanceHigh − trailingDrawdown), locking at the $0
starting balance. The EOD curve rebuilds from daily P&L + fund events,
with off-book fee drift spread across days proportionally to reported
fees (validated against a platform-reported MLL to ~$0.10). Stored in
autoLiqThresholds — negative, or +$0.01 once locked — so dead detection,
dashboards, and the ruin model's startThreshold all use the real room.
Guards move from '> 0' to '!== 0'; 0 still means no floor. canTrade=false
accounts are skipped. daily_pnl gains a fees column for the weighting.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
- TopstepXClient (ProjectX Gateway): API-key auth, SignalR user hub,
Trade/search daily P&L, emulated OCO exits, books-invariant
withdrawal detection for $0-start accounts
- Platform-aware client pool, credentials (api_key column), and
credential form; $0-size account types match by prefix alone
- Drain stages: min-day ($5-$15 randomized) MNQ legs for the
configured days, then wait for the 50% withdrawal; stage pills
show 'Stage N – Drain'
- Account calendar projects the perfect path: future days book their
planned targets, simulated 50% withdrawals, payout marker
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
When an account type opts in, legs are placed as an OCO pair via
order/placeoco - the usual limit plus a stop: the first cycle day's stop
keeps the day's net floor at -$1,000, min-day trades stop at 2x their
target, and the day after any red day rides stopless to make it back. A
stop-out books the day: eligibility refuses the account once realized
P&L falls short of the legs fired (or is simply negative on an opted-in
account, which survives restarts). If placeoco is rejected the leg falls
back to the classic lone limit with a loud log rather than leave a
position without an exit.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Big days (> the $1,700 split threshold) must carve at least the account
type's minLegs from the catalog — pickLegAmount reserves the minimum
piece for each leg still owed, forcing a carve even once the residual
dips below the threshold, clamped to what the pieces allow. Days at or
below the threshold still trade as one leg. The smallest split piece
drops from $800 to $500. An older master without the field means no
minimum.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Like accounts in a batch previously carved legs with independent
Math.random, so near-identical day targets could split very differently.
Leg fractions are now drawn from groupLegFraction(prefix, tradeDate,
legIndex) — FNV-1a seeded like dailyBump — so accounts of the same
account type on the same leg of the same day split identically when
their remaining targets match, and proportionally close otherwise.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Account types with the catalog's new combineCmeHolidays flag are read
through a combined-day view: holiday-session P&L entries fold into the
next full trading day (computed from the federal-holiday rules, so any
year works), the day count matches how the firm records it (Apex: Sun/
Mon/Tue around a Monday holiday is ONE day), and on the later sessions
of the stretch the earlier sessions' booked P&L carries into realized —
so day targets, already-traded gates, target-hit status and the panel's
day P&L all treat the stretch as a single day. Off by default; older
masters that don't send the flag behave exactly as before.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
runTrade now draws each batch from a single account-type prefix group
(largest group first) instead of slicing across prefixes; a group's
remainder fires alone on a later tick. copyTrade fills open slots with
accounts matching the positioned accounts' prefix before mismatched ones.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
When the master marks an account type splitEnabled=false, big days trade
as one unsplit leg (the legacy one-and-done behavior); min-day bumps are
unaffected. Missing field — older master or pre-toggle cached catalog —
defaults to enabled.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Big days (> $1700) now trade as multiple separate legs of random size
(each >= $800, $5-rounded), resumable across pauses since the remaining
target is derived from realized PnL against a day-anchored target.
Min-day targets get a deterministic random $0-25 bump seeded per
(account, trade date) so the dashboard, reporter, and orders agree.
Strict-consistency accounts randomize only their first trade ever.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
This box no longer decides what firms exist. It caches the master's catalog
and stores one thing of its own: the Tradovate login for each firm. Creating
firms locally is gone — app/add, POST /api/firms, the account-config routes
and the banned-symbol route with it.
The cache is the point, not an optimisation. The scheduler needs every
account's target and size on each 60s tick, so fetching them per tick would
stop trading with positions open the first time the master rebooted. It is
one JSON blob in settings, replaced only by a complete valid response, and a
response with zero firms is refused outright: a misconfigured FIRMS_DB serves
{"firms":[]} as a healthy 200, and mirroring that would idle the whole fleet
at once.
Accounts bind to an account type by prefix AND size, the size taken from the
opening funding transaction. That does not arrive with the first sync, so an
account sits Unassigned and the resolver tries again next sync rather than
guessing. A type picked by hand is never overwritten by an auto pass. This
replaces the longest-prefix helper that had been copy-pasted into five files,
which could not tell two types with the same prefix apart at all.
firm_credentials and account_assignments deliberately have no foreign key to
the cached catalog. foreign_keys is ON, so a cascade would erase the Tradovate
passwords the moment a firm was removed on the master — routine, not
exceptional. They survive it, and are re-keyed by name if the master's database
is ever rebuilt with fresh ids.
The legacy firms/account_configs/firm_banned_symbols tables are left on disk,
unreferenced. firms is the only table that ever held plaintext credentials and
dropping it would make a rollback unrecoverable.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Instances were started by hand and updated by hand, so they drifted behind
master silently. Now: PM2 supervises the dashboard and the clicker, a logon
task brings them up, and a 5-minute task pulls, rebuilds and restarts when
master moves.
Restarting on every push is only safe because the scheduler now survives it.
It was pure in-memory state (_global.__autoTrader), so any restart silently
stopped automated trading with the dashboard simply showing it as off. It now
mirrors running/action/symbol/stopAfterAll to the settings table, and
resumeSchedulerIfPersisted() picks it back up from the getClients() bootstrap.
No sync-wait was needed there: tick() already skips while a client reports
!syncComplete and while any account holds a position.
A failed build is never deployed — the build runs before anything restarts, so
a broken push leaves the previous build serving.
start-all and update-check both warm the app with a request afterwards. That is
load-bearing: getClients() is lazily bootstrapped, so until something makes an
HTTP request the Tradovate clients, the reporter and the resumed schedule never
start. That was already true of manual restarts.
Logic lives in Node so a macOS or Linux port only needs an equivalent of
install-autostart.ps1. Python deps are hash-guarded, so the common path is one
hash and one import with no network, and failure is non-fatal since only the
clicker needs them.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
`npm run build` failed on a clean checkout, so nothing on master could be
built for production. `npm run dev` does not hard-fail on type errors, which
is why it went unnoticed.
- state route returned client.perContractFees, which has never existed on
TradovateClient on any branch; nothing consumed it
- mapFirmConfig omitted bannedSymbols. Type gap only: the trade path calls
isSymbolBanned() against the DB directly, so bans were always enforced
- initClient's sync callback was sync where the constructor wants
() => Promise<void>
- accessInfo and ws are assigned during async connect/auth, never in the
constructor, so they take definite-assignment assertions
- the socket payload's inline entityType union had drifted five members
behind the indirect-callback union above it, making the 'position' and
'cashBalance' branches unreachable to the compiler. Both now share a
TradovateEntityType alias. Type-only: those handlers ran fine at runtime
Behaviour is unchanged throughout.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
The symbol resolution loop iterated every firm × every enabled symbol
calling findFrontMonthContract (Tradovate + Yahoo APIs), causing the
endpoint to hang for 30+ seconds with no logs. Now only checks the
client that actually holds the position.
Added logs at function entry and after each major resolution step so
hangs are visible in stdout.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Surfaces exactly which gate caught each ineligible account during
copy-to-max so we can debug which filter is excluding accounts.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
computeDailyTarget now requires equityProfit (amount - accountSize) and
returns null when it's undefined/null/NaN. 0 is still a valid value.
- Removed totalProfit parameter (was only used as fallback)
- Callers handle null by skipping the account (eligibility) or throwing
(execution paths)
- State API sets dailyTarget to null when no valid balance, avoids
incorrect targetHit computation
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Instead of relying on dailyPnL sum (which can miss reports or be out
of sync with account balance), use (amount - accountSize) as the actual
profit when comparing against profitTarget. Consistency calc still uses
dailyPnL totalProfit for realTarget.
Also unconditionally floor the min-day reservation at minDayPnL — if
equity + (days × minDay) >= target, we coast on min-day.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
The POST handler calls runTrade() directly, bypassing the scheduler
tick's position gate. If you stop and restart with open positions,
other accounts could get new trades while existing ones haven't closed.
Now runTrade itself checks for open positions before proceeding.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Replace volume-based contract selection with Yahoo Finance continuous
contract price matching ({PRODUCT}=F). The old approach compared volumes
across front/roll candidates, which failed when serial months had
deceptive volume (6EJ26 > 6EM26) or Yahoo was rate-limited (all 0s).
Now fetches the continuous contract price and matches it against
candidates within 0.1% tolerance. Falls back to roll1 if Yahoo fails.
Also adds User-Agent header to avoid 429 rate limiting.
Verified: GC=F price matches GCM26, 6E=F price matches 6EM26.
Also fixes stale 'Random' comment in auto-trade.ts and cleans up
frontVolume/rolledVolume references from settings page.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Instead of relying on the scheduler state (which may be 'Auto'), look up
each enabled symbol's front-month contract and match by contractId to
determine which symbol the positioned accounts are actually trading.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Bug fixes:
- Fix computeDailyTarget when consistency is 0% or 100%: treat as no constraint,
letting min-day reservation or full remaining profit drive the target
- Rename 'Random' to 'Auto' across entire codebase (types, API, UI, scheduler)
Features:
- Add "Stop after all eligible" checkbox: auto-stops scheduler when all
configured accounts are dead, inactive, already traded, or challenge complete
- Show position direction in status pill: "Long" (green) / "Short" (red)
instead of generic "In Trade" (blue)
- Add "Copy to Max" button: copies current trade direction to remaining
eligible accounts up to max_concurrent_accounts limit
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
- Add withdrawal stage system: each stage defines profit target, consistency,
and min trading days for post-withdrawal challenge cycles
- Target Same Equity mode accounts for withdrawn amounts when computing
effective profit target (profitTarget - remainingProfit)
- Store fund transaction timestamps for time-aware cycle filtering
(withdrawals before 9 AM CT include that day in new cycle)
- Expose full P&L history (fullDailyPnL) for calendar/equity curve display
across all cycles, with DB fallback for pre-restart data
- Show stage number (#1, #2, etc.) on calendar cells
- Hide consistency reference line when consistency is 0% or 100%
- Settings UI: "After First W/D" column with same-equity checkbox,
expandable stage sub-rows with profit/consistency/days inputs
- Default target_same_equity to 1 for new and existing account configs
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Dropdown on settings page with two options:
- Full CME (5:00 PM – 3:00 PM CT) with 5 min buffer
- Equity Hours (8:30 AM – 3:00 PM CT) with 5 min buffer
Setting is read live each tick, no restart needed.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Gives a full hour buffer before the CME daily halt at 4 PM.
No-trade window is now 3:00 PM – 5:00 PM Central (Mon–Thu),
and after 3:00 PM on Fridays.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
- Replace the reactive throttle with a proactive setInterval in requestSync
that fires fetchDaysTraded exactly once per hour per client
- Remove post-fill fetchDaysTraded calls from auto-trade.ts — no longer
needed and were causing bursts of report API requests on simultaneous fills
- Guard against duplicate intervals if requestSync fires more than once
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
- Seed tick_interval_seconds setting (default 60s)
- Expose tick_interval_seconds via GET/PATCH /api/settings
- startScheduler reads the setting at start time; enforces 5s minimum
- getSchedulerStatus returns intervalSeconds for the UI
- Main page: "Every [__] s" input in idle bar — saves on blur, persists across
restarts; running state displays "every Ns" next to symbol/action
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
- Symbol dropdown now includes a "Random" option alongside enabled instruments
- runTrade resolves 'Random' to a random enabled instrument once per batch,
so all accounts in the same tick trade the same symbol
- Import getInstruments in auto-trade.ts to support the resolution
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
- lib/db.ts: new firm_banned_symbols table with getBannedSymbols,
isSymbolBanned, and setBannedSymbol helpers
- app/api/firms/[id]/banned-symbols/route.ts: GET lists banned symbols,
PATCH toggles a ban for a given symbol
- app/api/firms/route.ts: include bannedSymbols[] in firm list response
- app/firms/[id]/settings/page.tsx: Instruments section shows all
globally-enabled symbols with a red toggle to ban/unban; banned
symbols display a "SYMBOL BANNED" pill next to their name
- app/page.tsx: FirmRows shows "ES BANNED" (or current symbol) pill next
to the firm name when the selected trade symbol is banned for that firm
- lib/auto-trade.ts: skip firms entirely when the trade symbol is banned
- types.ts: add bannedSymbols field to FirmConfig
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
All accounts in the same batch trade the same resolved direction.
Each new batch (after positions are flat) picks a fresh random direction.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
- Replace fill/ldeps and fill/list approaches with Tradovate reports API
- Add bearer auth to getreport polling (root cause of prior 404s)
- Use endDate = tomorrow to ensure current-session fills are included
- Count all traded days when minDayPnL is 0, otherwise count days >= minDayPnL
- Add PATCH /api/debug endpoint for proxying raw Tradovate API calls
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
isInNoTradeWindow() checks the current hour in America/Chicago (handles
CST/CDT automatically) and returns true from 15:00–16:59. runTrade()
returns early if the window is active, blocking all entries regardless of
whether the scheduler is running or a manual trade is triggered.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
- Privacy button: masks account names beyond the first 5 chars with bullets;
eye/eye-off icon toggles the mode in the header toolbar
- computeDailyTarget: accepts minDayPnL + minTradingDays params; when a
positive min floor is set and mandatory days remain, reserves future-day
profit so each day hits the floor (cap = remaining - futureReserve, floor
= minDayPnL); returns effectiveMinDay directly once profit target is met
but days are not yet satisfied
- auto-trade: passes minDayPnL/minTradingDays to computeDailyTarget; for
zero-floor accounts that have met the profit target but still owe trading
days, trades 1 MNQ in-and-out at market (extra-day mode) and bypasses the
normal target=0 skip gate via isMnqExtraDay flag
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
- DB migration: ALTER TABLE account_configs ADD COLUMN max_position_size INTEGER NOT NULL DEFAULT 0
- Added max_position_size to AccountConfigRow, createAccountConfig, updateAccountConfig in lib/db.ts
- Added maxPositionSize to AccountConfig type in types.ts (0 = no limit)
- GET /api/firms/[id] now returns maxPositionSize per account
- POST /api/firms/[id]/accounts and PUT /api/account-configs/[id] accept maxPositionSize
- Firm settings page: new Max Contracts column (blank = no limit)
- auto-trade: contracts capped at maxPositionSize when > 0
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
- Auto-trade scheduler fires every 60s; uses Promise.allSettled batch so no new trades fire while any position from the current batch is open
- Commission gross-up: read entryCommission from cash.realizedPnL after fill (fallback 2.5×contracts), grossTarget = target + 2×entryCommission
- Sync gate: TradovateClient.syncComplete flag; scheduler skips tick until every client finishes initial position/balance sync
- Contracts formula changed to Math.ceil so $1500 target = 2 contracts
- Removed all fee caching (perContractFees, recentFills, fillFee handler) from tradovate-class.ts
- Removed firm_fees table, getFirmFees, upsertFirmFee from db.ts
- Deleted instrument-configs API routes; removed Fees UI from firm settings page
- /api/instruments returns full {symbol, enabled}[] objects; dashboard filters to enabled-only for trade selector
- Added auto-trade, debug, orders, settings, and trade API routes
- Instrument selector on dashboard now driven by enabled instruments from DB
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
