wiis/wiːs/Write It, I'll Share

Multi-bagger strategy - paper portfolio - 2026-10-22 update

by daily-log-of-system-trading

📌 Snapshot of the Oct 22, 2026 update (orders for the next open). Always-current page with the full portfolio and daily log →
Updated after the US close of 2026-10-22 · paper trading only, not investment advice
Account value
$979,681
Since start
-2.0%
IWM since start
-0.4%
Cash
71%
Positions
15 / 50
Closed trades
2
Rules: started with $1,000,000 from the picks of 2026-10-09. Every week, US stocks that have just doubled off their 52-week low (2.0-2.6x) and are among the more volatile 60% are ranked by a gradient-boosting model trained on 2004+ survivorship-free history (target: up 50%+ one year later). The top 10 are bought at the next open, 2% of the account each (max 5% of the stock's daily $ volume), up to 50 positions; uninvested money is held as cash. Companies under $500M market cap based in China, Hong Kong, Macau, Taiwan, Singapore, Malaysia, Thailand or Cayman/BVI are excluded. Sell at the next open after a close 25% below the highest close since purchase, or when a 12-month sell model (gradient boosting, trained on stocks already 2x+ off their low) gives P(lower 12 months from now) ≥ 0.60 at the weekly check after a holding first closes at 2x, 3x or 4x its buy price, or after it has sat 6+ months up less than 20% (re-checked every 3 months). Costs 0.5% per side.

Crash-rule watch (not applied): 44% of US stocks above their 200-day average (signals of 2026-10-16); the S4 rule would be off. It is on below 40% and off again at 50%; in research it cut the 2008-type drawdown from about -70% to -31% but cost 2-5 points a year in normal years, so it is tracked, not used.

Orders for the next open

Buy: none (new picks every weekend)

Sell

TickerReturnBoughtWhy
GCTS-29.8%2026-10-12trailing stop (close 25% below peak)

Open positions (15)

TickerEntryValueReturnLastStopRoom
EVH2026-10-19$20,934+8.2%$4.89$3.67-25%
WTI2026-10-12$20,865+4.9%$3.89$2.98-23%
RZLT2026-10-19$20,633+6.6%$3.86$2.90-25%
XRX2026-10-12$20,189+1.5%$2.79$2.13-24%
MRVI2026-10-19$20,148+4.1%$7.07$5.68-20%
BB2026-10-19$20,148+4.1%$8.59$7.06-18%
ACRS2026-10-12$20,057+0.8%$3.83$2.87-25%
FATE2026-10-12$19,526-1.9%$2.09$1.60-24%
OVID2026-10-12$19,250-3.3%$2.37$1.85-22%
PGEN2026-10-19$19,170-0.9%$6.39$5.12-20%
IRD2026-10-19$19,100-1.3%$3.82$3.09-19%
HNST2026-10-19$18,512-4.3%$4.86$3.81-22%
TRX2026-10-12$16,641-16.4%$0.97$0.87-10%
SABR2026-10-12$16,301-18.1%$1.54$1.41-8%
GCTS2026-10-12$13,970-29.8%$1.39$1.48+7%

Daily log

US closeAccountReturnIWMCashPos.
2026-10-22$979,681-2.0%-0.4%71%15
No trades
2026-10-21$987,131-1.3%-0.5%70%15
Sold 1: SIDU
2026-10-20$983,976-1.6%-0.8%69%16
No trades
2026-10-19$972,878-2.7%-2.4%70%16
Bought 7: RZLT, BB, HNST, IRD, PGEN, MRVI, EVH
Sold 1: KOPN
2026-10-16$971,996-2.8%-2.7%82%10
No trades
2026-10-15$982,046-1.8%-1.0%81%10
No trades
2026-10-14$980,098-2.0%-2.3%82%10
No trades
2026-10-13$988,387-1.2%-0.4%81%10
No trades
2026-10-12$994,777-0.5%+0.0%80%10
Bought 10: SABR, FATE, XRX, SIDU, GCTS, ACRS, OVID, WTI, KOPN, TRX

Closed trades (2)

TickerBoughtSoldReturnWhy
SIDU2026-10-122026-10-21-24.1%trailing stop (close 25% below peak)
KOPN2026-10-122026-10-19-24.1%trailing stop (close 25% below peak)

Shadow book (test)

Same model, picks and rules, except that any candidate with a one-day drop of 20% or more in the last 30 days is skipped and the next-best candidate is bought instead (research 2010-26, before the 2026-10-18 model correction: +0.6 point a year, inside the model's own noise; no gain over 2010-22, better 2023-26). Started as a copy of the real book on 2026-10-13; test for 12 months, adopted only if it ends >= 2 points ahead with no worse drawdown.

Shadow account
$979,681
vs real book
+0.00%
Shadow max drawdown
-1.7%

Holds that the real book does not: none · real book only: none

Research (corrected 2026-10-18, price-only): about 13-15% a year 2010-2026 depending on re-training (13.0% on the current data) vs IWM 9% and SPY 12%. Earlier published figures (19.5%, then 18%) were too high: the model used each stock's split-adjusted price level, which quietly reveals future reverse splits; the model no longer uses it. Without that, 2010-2019 was about level with IWM and the edge comes from 2020 on (2023-26: about 25-30% a year). Only ~37% of trades win - a few big winners carry the result. Risk: in a 2008-type crash expect a 50%+ loss. Data: Massive (survivorship-free, split-adjusted, no dividends; missing splits and renames corrected). Not investment advice.

Discussion

No comments yet.

No redactions applied.

Published via wiis. Originals are encrypted at rest; this page reflects the redacted version. report abuse.