8
Validated systems
2
Validated w/ caveats
9
Mechanical patterns FAILED
7
Walk-forward windows (Grimes)
Methodology: All systems validated via 7-window walk-forward (1999-2026). Grimes binding amendment requires realistic-mode (next-open execution) Exp ≥ 1.0 before any live deployment. System W realistic mode FAILED until partial-TP 2R/3R was added — this is the disclosure on every paid Substack issue.

Validated Systems Scoreboard

Status: GOOD = ready post-validation · WARN = needs tuning before live · BAD = retired

SystemSame-day %Realistic %WRExpMaxDDN tradesWalk-forwardStatus
System W (US)
Trend / momentum · Webster + O'Neil
+60731% +288% 54.0% 4.88 13.8% ~600 7/7 sameday, 5/7 realistic BEST CANDIDATE — partial-TP 2R/3R locks Exp >1.0
System DI (US)
DI+W+TML · current_branch
+26697% -33% 52.0% 1.50 35.0% ~700 7/7 sameday, FAIL realistic US gap-risk problem; realistic mode needs exec tuning
System W (Thai port)
Trend / momentum · System W
+326% 71.8% 13.8% ~120 single-pass, walkforward pending Cross-market PASS — methodology not US-only
First-Pullback (US)
Pullback after breakout · Komar/Webster/Kell
+8% 70.8% 89 3/7 overall, 3/3 RECENT Strongest 2020s edge per backtest
Failed-Reentry (US)
Re-entry after failed breakout · Oliver Kell #5
+4% 60.0% 13087 stable EVERY decade incl 2020s Statistically robust — wide N
Consolidation Breakout (US)
Long base breakout · O'Neil
+6% 63.5% 9953 5/7, FAIL window 7 (2020-2026) Modern era degraded in US, holds in Thai
Consolidation Breakout (Thai)
Long base breakout · O'Neil
+5% 52.0% 9 small N but 100% WR tight≤8%/vol≥20× Niche modern Thai edge
Climax-Top Scanner
Exit signal · O'Neil Sell Rule #4
score 65=33.8% TP rate (modest edge) Live in scan_climax_live.py
PTJ 200-day veto
Filter overlay · PTJ
-29% +53% 1.5% 9.00 WR +1.5pp, Exp +9% on existing systems VALIDATED — live in scan_di_live*.py
Partial-TP 2R/3R
Exit overlay · Minervini
-3% 13.8% 4.88 -1.6% Flips W realistic FAIL→PASS VALIDATED — Grimes blocker resolved

The Honest Failure List — 9 Mechanical Patterns That Failed Validation

Published openly. The hover is honest.

PatternTypeLineageFailure mode
VCP (mechanical) Pattern Minervini-mechanical 0 big winners over 36yr
Cup-with-Handle (generic) Pattern O'Neil-mechanical 0 big winners; named patterns over-specify
Shelf-Breakout Pattern Webster 22 trades 36yr, 0 big winners
Holy Grail (Raschke) Re-entry Raschke -80.7% return, 85% MaxDD despite 1.5R Exp
Komar Blue Dot filter Filter Komar Contradicts breakout entry; collapses N from 1015→7
Bonde Momentum Burst Pattern Stockbee Blew up -100% despite 53% WR
Turtle Soup as W filter Filter Raschke Cut total return 96% (over-filters strongest signals)
ACD (daily) Pattern Mark Fisher Designed for futures; daily stocks doesn't translate
Hunter Confluence Composite In-house 86,416-signal backfill showed all classifications inverted

3-Branch Backtest Allocation Plan

BranchPrimary systemPaper allocationValidation gateLive decision
simons_branch Ensemble (consolidation + first-pullback + failed-reentry) ฿333,333 / $33,333 USD 5d-hold equal-weight Sharpe ≥ 0.5 in walk-forward Best Sharpe at Nov 2026 wins live capital
oneil_branch CANSLIM C+A+L+I + Cup-with-Handle ฿333,333 / $33,333 USD All 8 sell rules enforced + ≥50% WR over 6mo paper Best Sharpe at Nov 2026 wins live capital
current_branch System DI + System W (existing) ฿333,333 / $33,333 USD FROZEN — paper-track baseline only Best Sharpe at Nov 2026 wins live capital

DCA Truth Scorecard (Thai)

15-year DCA backtest, 180 stocks, dividends 10% WHT, 18 corp-action stocks excluded

73
Winners (40.6%)
107
Losers (59.4%)
−6.4%
Median total return
+1024%
TISCO DRIP example (16yr)

Top 10 Verified Compounders (Cash-Only DRIP, 16yr)

#SymSectorFinal $ (from 100K)Total ReturnCAGRDRIPs
1AOTAirports3,379,016+3,279%24.0%15
2PTTEnergy2,458,864+2,359%21.2%32
3AIIndustrial1,993,113+1,893%20.5%22
4TISCOBank1,720,869+1,621%18.4%20
5ADVANCTelecom868,309+768%13.8%35
6BOLTech Services664,873+565%12.0%29
7CCETElectronics441,778+342%9.3%30
8ASIANConsumer430,038+330%15.7%16
9AMARCCommercial229,681+130%31.6%8
10BANPUEnergy47,752−52%−4.5%30
BANPU's apparent "winning" came from stock dividends. Cash-only DRIP shows the truth: −52% in 16 years.
Research-not-advice. Past performance does not guarantee future results. System W realistic-mode FAILED Exp 0.86 before partial-TP 2R/3R fix. Disclosed in every paid Substack issue. 6-month paper period sacred — first live trade November 1, 2026.

Generated 2026-05-16T16:30:25.685212+07:00