⚠️ Personal research and trading journal — not investment advice. The author does not provide licensed advisory services.
I ran an experiment that surprised me.
I took the same trades — same entry dates, same stocks, same position sizes — and applied four different exit strategies. The results ranged from +1R cumulative to -17.4R cumulative.
Same entries. Four exit strategies. Spread of 18.4R.
The Experiment Setup
My paper trading system generates entry signals from a breakout scanner. Once a stock enters, there are four different ways I could manage the exit:
Pattern branch (what I use): Exit based on a specific technical structure derived from the original chart pattern. For the first_pullback system, this uses the Webster point exit — price closes below the specific pivot structure that validated the entry. For failed_reentry, it uses a partial-TP with 21-day MA trail.
O'Neil branch: Apply O'Neil's 8 sell rules. This includes: cut losses at 7-8%, sell into 20-25% gains, sell when stock breaks the 10-week MA with heavy volume, sell if stock extends 70-100% from base. This is the approach from IBD/O'Neil's published methodology.
IBD simulation (ibd_sim): Exit when IBD commentary flags the stock as "extended" or gives a sell signal — simulating following IBD's own sell timing.
Current branch: Simple rules: 7% stop + 21-EMA trail.
The Results
Across all entries in my paper period (approximately 6 months):
| Exit strategy | Cumulative Managed R |
|---|---|
| Pattern (Webster PT) | +1.0R |
| Current (7% stop + EMA trail) | -9.3R |
| O'Neil (8 sell rules) | -1.0R |
| IBD simulation | -17.4R |
The pattern branch was the only strategy that produced net positive results. The IBD simulation was dramatically negative.
Why the IBD Simulation Failed So Badly
The -17.4R from following IBD sell timing reflects a structural problem: IBD commentary on selling typically lags price action.
When IBD says "extended, consider taking profits" — the stock has already run 30-40% from breakout. By the time the sell signal appears in IBD commentary, momentum has already peaked. If you hold until IBD tells you to sell, you're selling into declining momentum with much of the gain already given back.
This aligns with the IBD buy-zone finding from my earlier research (see "IBD Said Buy. 91% Moved Higher. Most Gave It All Back."): IBD's timing signals are systematically late for both entries (buy-zone means the optimal entry passed) and exits (sell signals arrive after the peak).
Why O'Neil Rules Underperformed
O'Neil's rules are more aggressive about taking gains — sell 20-25% — which should help with the "gave it all back" problem. But they also generate earlier exits on winners that then extended further. The 7-8% stop is tighter than my typical structure stop, so the rules also stopped out more trades that would have recovered.
The net effect: the O'Neil rules performed better than the IBD simulation (-1.0R vs -17.4R), but still underperformed the pattern-specific exit by 2R.
Why the Pattern Exit Outperformed
The Webster Point exit uses the specific chart structure to define when the trade is invalidated. If I entered on a pullback to a prior breakout level, the exit is defined by what the chart says the trade requires to remain valid: price holding above that structure.
This means: 1. The exit is anchored to the reason for the trade, not an arbitrary time or percentage 2. The exit only triggers when the underlying thesis breaks — not on normal volatility 3. Winners that extend past expectations continue running without an arbitrary cap
The weakness of this approach: when trades go wrong, losses can exceed what a tight 7-8% rule would produce. The trade-off is higher expectancy per winner at the cost of more variance per loser.
The Broader Principle
Most trading discussion focuses on entry: when to buy, what pattern to look for, which scanner to use. The exit is treated as secondary — "just cut at 7% and target 20%."
The branch replay experiment shows that's backwards. When you give the same entries to four different exit strategies, you get results ranging from deeply negative to positive. The entry is the same. The exit is what differentiates.
This doesn't mean entries don't matter — they do. But it suggests that matching your exit strategy to your entry strategy is as important as the entry selection itself.
The O'Neil approach has a coherent theory about exits. The IBD simulation doesn't have a systematic exit — it follows commentary that lags. The pattern exit has exits derived from the same chart logic that generated the entry.
Internal coherence between entry and exit is the principle. When you use pattern-based entries, use pattern-based exits. When you use trend-following entries, use trend-following exits. Mixing frameworks is where negative R accumulates.
Track. Study. Wait. Strike.
Personal research and trading journal — not investment advice. The author does not provide licensed advisory services. — MOEasymmetry
Draft 2026-06-12. Source: replay_branch_journal.py — applies each branch's own sell rules to paper journal entries. 4 branches: pattern (Webster PT), O'Neil (8 sell rules), ibd_sim (IBD commentary timing), current (7% stop + EMA21 trail). Cumulative R: +1 / -1 / -17.4 / -9.3. Same entries, same position sizes; only exit differs. Period: live paper journal (approximately 6 months). Sample size is small — treat as directional signal, not conclusive evidence.