How to Compare Trading Strategies (Metrics & Process)

Every active trader eventually runs more than one idea: an opening range breakout, a mean-reversion scalp, a swing pullback. Without a comparison framework, capital and attention drift toward whatever worked last fortnight. That recency bias destroys edges that need time to prove themselves and keeps alive setups that already failed statistically. Comparing trading strategies properly means isolating each playbook, aligning time windows and costs, and reading a small set of robust metrics — expectancy, profit factor, drawdown, and sample integrity — instead of debating win rate at the lunch table. This guide is a complete playbook for that comparison process inside TradeLyser and in your own risk policy.

Why strategy comparison is not optional

Blended account P&L answers “How am I doing?” It does not answer “Which behaviour pays?” A green month can hide a broken intraday system subsidised by one swing winner. A red month can hide a solid setup drowned by untagged impulsive trades. Per-strategy analytics turn capital allocation into an evidence exercise: more size on stable positive expectancy, pause or shrink size on deteriorating curves, kill playbooks that never earned their sample.

Define the strategy before you measure it

A strategy document should fit on one page: market (cash/F&O), instruments, session window, entry trigger, invalidation, target style, max risk per trade, and max daily risk. If two setups share entries but different exits, they are two strategies. If you cannot explain the difference to another trader in thirty seconds, tagging will drift and metrics will lie.

In TradeLyser, create strategies before bulk tagging historical trades. Retro-tagging is fine for a pilot window, but avoid retagging old trades under new names mid-comparison — that rewrites history.

Prerequisites: when comparison is valid

• Each strategy has at least 30–50 closed trades (100+ for slower swing or complex options structures).
• Tags are mutually exclusive per trade unless you explicitly model overlap rules.
• Same calculation method (FIFO vs average) for the entire window.
• Costs, taxes, and slippage reflected in net P&L settings.
• Comparison window is identical calendar span for all candidates.

Core metrics and how to read them together

Expectancy

Expectancy is average rupees or R-multiple per trade after costs. It is the single best “repeat this 100 times” number. Positive expectancy with acceptable drawdown warrants continued research; negative expectancy with huge win rate means losers overwhelm winners — a classic trap.

Profit factor

Gross profits divided by gross losses. Below 1.0 is net losing. Many disciplined systems live between 1.2 and 1.8 depending on style. Spikes above 2.0 on small samples often mean one outlier week — verify distribution, not just the ratio.

Max drawdown and recovery

Drawdown tells you capital and psychology required to run the playbook. Two strategies with similar net P&L can have vastly different pain profiles. Pre-write max drawdown pause rules per strategy in rupees or percent of allocated capital.

Win rate (context only)

Win rate without average win/loss size is marketing. Use it with payoff ratio and expectancy. See our win rate analytics guide and glossary entry for depth.

Eight-step comparison process

• Pick a fixed window (e.g. last 90 calendar days or last 200 trades per strategy).
• Export or open side-by-side strategy stats in Strategy Board.
• Record trade count, net P&L, expectancy, profit factor, max drawdown for each.
• Slice by symbol — flag if one ticker drives >40% of profits.
• Slice by time of day — flag if edge is only first 45 minutes.
• Check rolling 20-trade expectancy trend — improving, flat, or decaying?
• Compare to your written risk limits — any strategy breaching pre-defined drawdown?
• Decide capital allocation changes in writing before Monday open.

Capital allocation after comparison

Think in buckets: core (proven expectancy, stable drawdown), research (promising but insufficient sample), and paused (negative expectancy or rule chaos). Cap research bucket risk low. Never promote a strategy to core because of one viral week on social media.

Example policy: core strategies share 80% of risk budget split by rolling expectancy weights; research strategies share 15%; 5% cash buffer for mental clarity. Numbers are personal — write yours.

Overlapping strategies and tagging hygiene

If one trade could be two tags, pick a primary tag rule: first touch, dominant thesis, or largest risk. Document the rule. Double-tagging without accounting splits metrics and inflates trade counts artificially.

Market regime changes

Low volatility, high volatility, trend, chop — edges rotate. Rolling metrics detect decay earlier than monthly P&L. When rolling expectancy turns negative while lifetime stats look fine, treat it as a regime alert: reduce size, paper trade updates, or pause until sample rebuilds.

Psychology of killing a beloved setup

Traders anchor to narratives (“This is my bread and butter”). Data should trigger investigation, not instant deletion. Protocol: two consecutive review cycles below expectancy floor → pause → 20-trade paper sample with one rule change → resume only if paper metrics recover. Emotional attachment is why written policies exist.

Using TradeLyser Strategy Board

Create and archive strategies in docs-linked how-tos; this article is the analytical layer. Use compare views for side-by-side tables, track performance over time for equity curves per strategy, and analyse stats for distribution detail. Link journal notes on weeks you changed rules so future comparisons remember context.

Comparison mistakes

• Different date ranges per strategy.
• Including open trades in closed-trade metrics.
• Ignoring partial fills and corporate actions on equities.
• Scaling size mid-window without noting the date in journal.
• Comparing a discretionary “feel” bucket with a mechanical setup — rename or split.

Comparing options strategies separately

Iron condors, naked shorts, and directional long options have different win-rate profiles. Never blend under “options”. Tag structure type. Compare max loss trades explicitly — one gap morning can define monthly drawdown for short premium books.

Documentation traders should keep outside the app

Maintain a one-page strategy spec PDF per playbook: charts examples, invalidation, size formula. Link in notebook. TradeLyser holds performance; your spec holds logic. Update version numbers when rules change and note version in journal for that week.

If you run multiple accounts or family pools

Compare per account, not blended, when risk policies differ. Switch-account features in docs support separation. Allocation policy should not merge conservative IRA-style logic with aggressive intraday wallet.

Statistical humility

When difference between two strategies is small, assume noise until sample grows. Use confidence language: “B leads A on expectancy by 0.3R over 80 trades — monitor, not finalize.” Avoid permanent kills on 12-trade samples.

Closing

Strategy comparison is how you stop being a collector of setups and become a portfolio of edges. Run the eight-step process monthly, enforce allocation policy weekly, and let journal discipline protect you when metrics lag. Open TradeLyser, pick two strategies you run today, run the comparison tonight, and write one allocation decision before the next session.

When backtesting historical trades imported via CSV, tag backtest separately from live — never blend expectancy. Comparison is live-on-live, research-on-research. Mixing contaminates capital decisions.

Institutional desks run “playbook committees” monthly; you can run a one-person committee with printed metrics. Print side-by-side tables, mark decisions in ink, photograph the page, store in notebook. Digital metrics drift; ink decisions feel binding. Committee agenda: confirm sample sizes, confirm tagging audit spot-check of ten random trades, confirm drawdown versus policy, vote pause / continue / research for each playbook. No vote without written reason. Over a year, your committee minutes become the most honest trading biography you own — more truthful than social posts, more actionable than memory.

When capital is limited, run at most two live strategy slots plus one research slot. Slots are risk budgets, not ego badges. Empty fourth and fifth slots until evidence earns them — white space is discipline. Comparison meetings decide what fills white space next quarter, not what filled it last year on nostalgia.

Link comparison outcomes back to journal rules

Every allocation decision should appear in next week’s journal template as a risk rule: “Max 40% margin on Strategy A until sample hits 100 trades.” Journal enforces what comparison decided. Without journal linkage, comparison slides are PDFs in a folder while live trading reverts to old habits by Wednesday.

Extended matrix: scoring strategies for capital slots

Create a simple scorecard 1–5 on five axes: sample adequacy, expectancy stability, drawdown fit to your risk budget, behavioural ease (can you execute calmly?), and regime sensitivity. Sum scores; allocate capital slots top-down. Refresh quarterly. This prevents narrative favourites from hogging margin while statistically superior setups starve.

Document kill criteria in advance: three consecutive weeks below expectancy floor, or drawdown beyond slot limit, triggers pause — not debate. Paused strategies enter paper tracking for twenty trades before resurrection. The matrix makes kills procedural, not emotional.

When fundraising or explaining performance externally, never quote blended win rate — quote per-strategy expectancy with trade counts. Serious capital asks for distribution, not highlights.

Revisit comparison after major rule changes with a new sample window — old stats are invalid once entry criteria shift. Archive PDF of old rules with end date in notebook for forensic honesty.

Collaboration tip: if you trade with a partner, compare per trader tag within shared strategies to see whose execution drags expectancy — uncomfortable and useful.

Synthesis: comparing strategies is how you respect your own capital. Capital is time and effort converted into rupees; allocating it without comparison is guessing. The comparison framework in this guide is deliberately boring — fixed windows, joint metrics, written decisions — because boredom scales. Exciting decisions are tweets; boring decisions are desks that survive. TradeLyser Strategy Board makes boring faster, not optional. When you feel tempted to add a shiny new setup, run the comparison against existing playbooks first. If the new setup cannot beat something on expectancy after honest sample, it is a hobby, not a slot. Hobbies are fine outside market hours; they are expensive inside margin accounts. Write allocation policy when calm; execute when greedy voices appear. The voices will appear after wins. Policy beats voices. Review policy quarterly, not after every trade. This is how small accounts become medium accounts without heroic stories — just repeated, documented allocation calls that favour evidence.

Portfolio thinking with small accounts

Small accounts still benefit from strategy slots — even two strategies teach allocation thinking. Slot risk in rupees, not only percent, when absolute loss matters psychologically. A ₹5,000 daily max loss feels different at ₹2L versus ₹20L capital — write rupee limits.

Correlation between strategies matters: two mean-reversion Nifty scalps are one risk bucket, not two. Tag correlation family in notebook when comparing.

Scaling rules: increase size only after twenty trades post-rule-change with stable rolling expectancy — not after one week. Document size ladder in strategy spec.

Public track records often cherry-pick strategies. Your private comparison is the honest one — protect it with tagging discipline.

When Strategy Board shows a tie, prefer simpler strategy with lower operational error rate — fewer moving parts win long run.

Annual strategy retrospective: archive one-page summaries of each playbook with final trade count, net P&L, expectancy, and reason if paused. This archive becomes your institutional memory — invaluable when beginners ask what actually worked for you versus what you wish had worked.

Execution grade versus statistical expectancy

Document not only which strategy wins but which strategy you can execute calmly at scale. Calm execution beats theoretical expectancy you sabotage with hesitation or oversize. Comparison tables should add a subjective execution grade 1–5 each month — honest, private. Low execution grade with high expectancy means practice on simulator, not live size-up. High execution grade with low expectancy means pause, not push. The matrix of math plus execution keeps you from becoming a spreadsheet trader who cannot click buttons consistently under stress.

When markets shift regime, rerun comparison with last regime window versus prior regime window — side by side. Regime tables explain why a strategy “stopped working” without moral panic. Label each window with a one-line regime note — trending, range-bound, high VIX — so you do not confuse structural edge loss with temporary variance.

This flagship strategy guide anchors the strategies pillar — pair it with glossary entries on expectancy and profit factor and with TradeLyser Strategy Board after tags are clean.