Forums compare setups with opinions. Practice compares them with identical logging, caps, and hidden dates.
A/B practice blocks with identical logging. Same rules, different setup—let process scores decide.
A/B block design (2-week minimum)
Week A: Setup 1 only (e.g., pullback long)
Week B: Setup 2 only (e.g., breakout continuation)
Constants:
- Random hidden NSE dates
- Same session window focus
- Same entry cap (e.g., 3)
- Same risk unit in sim
- Same process rubric 0–2
Do not change constants mid-experiment.
Logging grid (both weeks)
| Field | Setup A | Setup B |
|---|---|---|
| Sessions run | ||
| Total decisions | ||
| A-grade % (of entries) | ||
| Cap violations | ||
| Avg process score | ||
| Skip quality (subjective audit) |
Avoid P&L as primary column—it tempts fake certainty. Outcome can be noted, not worshipped.
What comparison can and cannot tell you
Can suggest:
- Which setup you follow more cleanly
- Which condition tags pair with fewer rule breaks
- Which urge (fomo, boredom) appears more
Cannot prove:
- Live expectancy
- Win rate superiority
- "Best setup for Nifty" tip culture nonsense
NSE context for fair comparison
Compare on liquid names unless both setups target thin tape deliberately. Tag open vs midday sessions separately—breakouts may look better at open in sim while pullbacks behave differently midday.
Corporate-action-heavy random draws may distort one setup—note anomalies, do not hide them.
Failure modes
- Changing caps between weeks
- Known hero days in one week only
- Adding indicators in week B
- Declaring winner after 4 sessions
- Ignoring skip quality
After the block
Keep setup with higher process adherence and lower cap violations. Drop or defer the other—not forever, but until a future block with new condition focus.
Niche discovery uses the same matrix thinking: setup × condition × timeframe.
Replay Trader connection
Replay Trader supports setup comparison via randomised NSE replay and structured reps: ₹999 / 30 days, 300 decisions repetition hypothesis. No forum drama, no guaranteed profits, no stock tips.
Compare setups in practice ledgers, not comment threads. Educational simulation for Indian traders.
Statistical humility note
Even twenty sessions per setup is not rigorous A/B testing. Enough for personal process comparison, not publishing "Setup A wins."
Third setup deferral
If tempted to add Setup C mid-block, defer to next month. Comparison requires discipline too.
Condition fairness
Do not run Setup A on open week and Setup B only on known trend days—you biased the block.
Shared rubric document
Print process rubric once; both weeks use identical definitions of A-grade.
After comparison live rule
If live trading, pick higher-adherence setup for smaller size first—still no expectancy claim.
Blind week option
Log Setup A sessions without noting which setup until week ends—reduces midweek tinkering (advanced).
Shared market draw
Same random date both setups on different days unfair—different conditions OK if hidden.
Setup journal separation
Physical divider in notebook between weeks.
Comparison presentation
End month: table only, no tips to friends.
Next block scheduling
Schedule Setup C test month two—not immediately.
Extended two-week calendar
Week A Mon–Fri: Setup 1, open window, hidden dates, cap 3, min six decisions logged daily.
Week B Mon–Fri: Setup 2, same constants.
Weekend: compare tables, no rule changes allowed until comparison memo written.
Memo answers: Which setup had higher adherence? Which had cleaner skip reasons? Which paired with fewer urge tags? If answers tie, run third week tie-breaker with midday focus—not P&L tie-breaker.
Comparison integrity oath
No rule changes during two-week block. No peek. No hero dates. If oath broken, restart block clock. Comparison data worthless otherwise—worse than worthless because false confidence.
Extended A/B integrity checklist (daily)
Before each session during comparison block:
- Same cap as block charter
- Same risk unit
- Same scoring rubric
- Hidden date not re-rolled
- Setup matches week A or B assignment
- No new indicators
If any unchecked, session marks block_contaminated and does not enter comparison stats.
After block: write memo answering: Which setup did I execute cleaner? Which urges appeared more? Which condition cells differed?
Memo forbidden from containing win-rate brags or stock tips.
Comparison serves niche and fit decisions—not marketing claims.
Post-comparison decision tree
If Setup A cleaner on process → month two deepens Setup A in weakest condition cell.
If tie → third week tie-breaker with stricter caps, not P&L.
If both messy → neither live; return to patience and cap fundamentals.
Comparison without decision tree becomes forum argument archived in Notes app.
Comparison archive
Store weekly tables in one folder named setup_compare_YYYYMM. Future you compares month-over-month adherence, not month-over-month win rates. Archives also prevent re-running the same comparison while pretending it is new insight.
When block ends, delete temporary indicators added for curiosity—comparison integrity extends to tool hygiene.
One-line block charter example
"Week A pullback-long, open window only, cap 3, min 6 logged decisions, hidden dates, rubric v1 frozen." Pin charter above desk. Any session that violates charter before click starts is abandoned and restarted—comparison integrity begins at session start, not post-hoc excuse.
Related reading
Replay Trader — randomised historical NSE practice, hidden dates, candle-by-candle reveal — is built for decision reps under uncertainty, not tip culture.
Practise today. Trade tomorrow.
Educational content only. Not investment advice. Not a recommendation to buy or sell any security. Practice and simulated results do not guarantee live results.