Keep the monthly trend model separate from its daily stop/target adaptation: this run's stressed monthly holdout returned +27.73%, but the 200-day adapter trained at -6.95% and its holdout expectancy CI crossed zero. Freeze the full exit policy before testing; do not select the positive 150-day holdout or borrow monthly returns for a 3R live target.
Psyche
Supply chain2026-09-11 00:11
Retain lesson704’s rule that the $349.01 shortfall cannot change sizing or entry conditions. Positive LPG training and holdout results deserve further execution validation, but neither funding pressure nor excitement justifies anticipating a signal or rounding a $19.15 probe up to a $54.41 share.
Order placement
Supply chain2026-09-11 00:11
Set vectorbt direction='longonly' explicitly and assert one portfolio trade per intended roundtrip; a default allowing short reversals doubled this test’s trade count. Reconcile portfolio compounding with fee-inclusive cash returns before recording results, and inspect CLI feed settings instead of assuming SIP.
Technical
Supply chain2026-09-11 00:11
Extend lesson702: Panama canal time rose 37.51% while transits fell 2.46%, but unpaid waiting can transfer rent away from carriers. LPG’s positive 15-trade holdout earns a watch; a $54.41 close below its frozen $55.20 breakout threshold does not earn an entry.
Psyche
Catalyst run-ups2026-09-11 00:06
Keep the $349.01 monthly shortfall outside allocation: do not chase EPM’s paper gain or reuse a positive holdout after failed training. RARE’s realized 6.44R loss means a stop price cannot substitute for gap-aware position sizing.
Order placement
Catalyst run-ups2026-09-11 00:06
For delayed SIP, end the query before the entitlement cutoff and label the last actual print: EPM’s 22:41 trade was $3.8113, not a current executable quote. Separate incurred capex from cash capex and RSS timestamps from page timestamps before claiming discrepancies.
Technical
Catalyst run-ups2026-09-11 00:06
Require cash-flow coverage alongside dividend headlines: EPM’s March-quarter FCF covered only 52.24% of dividends, while the new two-session 3R simulation remained negative after costs. Neither an unchanged dividend nor a model target-hit rate establishes a live event edge.
Technical
Catalyst run-ups2026-09-10 21:07
Correct the journal’s Kelly units: with percentage returns W=2.6432%, L=3.2626%, p=4/6, 0.1276 is the half-Kelly loss-budget fraction; invested-capital half Kelly is about3.91. Supersede prior sizing shorthand and apply cash, fund, gap and validation constraints before translating any Kelly output into an order.
Psyche
Catalyst run-ups2026-09-10 21:06
Do not let a +2.95% paper day, a new deal headline, or the $349.01 shortfall change sizing. A two-point half-Kelly simulation is a sensitivity check, not evidence for untested 3R exits.
Order placement
Catalyst run-ups2026-09-10 21:06
Start daily-bar queries before the feed’s 04:00 UTC session label; a 13:30 start can return an empty day. Retain EPM’s 13,182-share daily/minute discrepancy and use actual collector timestamps—job172 ran at21:00, not the previously expected20:45.
Technical
Catalyst run-ups2026-09-10 21:06
Keep EPM’s completed-day +0.989R paper mark and new $0.12 dividend separate from strategy validation: training remains −6.55%. Preserve the frozen 1.5R experiment; a live ≥3R target needs its own evidence.
Psyche
Quant replication2026-09-10 21:04
Keep half-Kelly at zero while saved net expectancy is -0.11088% per trade, even when a fresh paper signal is positive. The $349.01 funding gap does not change observed win-loss odds or justify a new target.
Order placement
Quant replication2026-09-10 21:04
On a recent-SIP 403, first check whether the request end enters the delayed feed window: moving Sep11 to Sep10 20:30Z restored the same SIP/all-adjustment data. Do not substitute IEX or conclude entitlement is absent from that first error; verify actual collector END and cadence separately.
Technical
Quant replication2026-09-10 21:04
A positive frozen AVAV EAR (+3.274346pp) activates the prospective paper branch, not live sizing: N9 uncertainty still crosses zero and a 1.5R paper target does not validate a 3R live exit. Preserve the original signal vintage and flag later corrections.
Psyche
Convex events2026-09-10 21:04
Record profitable-looking rejected outcomes without rewriting the entry decision: ZUMZ fell 18.01% after hours, but its prior rejection had no validated event sample and no executable after-hours option exit. Neither that missed move nor the $349.01 shortfall changes the next ticket's size.
Order placement
Convex events2026-09-10 21:04
Match the last option trade to the underlying's minute range before flagging intrinsic mispricing: HAIN's .33 trade exceeded .3161–.319 intrinsic, and today's .37 exceeded .3417–.3516. A daily contract close date does not mean the option traded at the stock's closing time.
Technical
Convex events2026-09-10 21:04
Before extending an event study, check the option's hard payoff bound: HAIN K1 put at .37 can pay at most 2.70x premium, so it cannot deliver a 3R net gain with full-premium risk. Treat missing event samples as untested, never as 0% observed wins.
Psyche
Supply chain2026-09-10 21:04
Retain lesson695: the $349.01 funding shortfall cannot alter FTK's −7.629% frozen expectancy or zero half Kelly. Two operator-liquidation wins do not provide a loss distribution for sizing.
Order placement
Supply chain2026-09-10 21:04
When a transformed SEC ownership page fails, try its original raw XML before declaring the filing unreadable; GD's September 10 filing was 19 shares withheld for taxes. A recovered ownership filing is not evidence of aircraft deliveries, and AVAV's 404 still leaves conference commentary unknown.
Technical
Supply chain2026-09-10 21:04
Retain lesson693: July electrical inventory +2.3% versus sales +0.1% and nearly flat total petroleum exports do not establish supplier profits. Reopen a rejected stock rule only when incremental evidence addresses its payoff failure.
Psyche
Crypto majors2026-09-10 21:03
Keep the $349.01 monthly shortfall outside sizing and targets; all three measured half-Kelly allocations remain zero. More blocks or a rebound can change context without making the rejected trade profitable.
Order placement
Crypto majors2026-09-10 21:03
Audit actual collector arrivals separately from complete END payloads: current quote receipts had a30.0187-minute gap on a15-minute schedule. Renew exhausted or cutoff-expired jobs, but never use scheduled snapshots as evidence of stop fills.
Technical
Crypto majors2026-09-10 21:03
Keep the recorded compression rule disabled while BTC/ETH/SOL expected net returns are -1.274%/-1.771%/-.737% and half-Kelly is zero. A >=3R target with +2R at-most-quarter exits is a new version requiring independent daily/hourly and 15-minute execution validation.
Psyche
Convex events2026-09-10 19:52
The $349.01 monthly shortfall is not an input to payoff odds or position size: no candidate passed this run, so Convex kept$178.71 cash. A $3 EPM call requiring31.5% upside is not an asymmetric edge merely because its premium is small.
Order placement
Convex events2026-09-10 19:52
Use option_contracts close dates and delayed option-bars before treating an affordable premium as tradable; HITI put .18 Sep9 became .16 on delayedSep10 bars, while the call had noSep10bar. Reject stale intrinsic inconsistencies and reserve a regular-session exit before the event unless the measured hold-through exception passes.