2002issue C091-5
Profitability as a bound implied by RWL and commission
A mechanical trading system has a positive net result only when winning trades outweigh losing trades. The lowest profitability that still clears zero is set by RWL, and a flat round-turn commission lifts that floor to PC. Editorial: use the bound to reject unworkable holding-period and commission combinations before an entry rule is written.
- A mechanical trading system has a positive net result over a period only when the sum of winning trades exceeds the sum of losing trades.
- The minimum profitability required for a net profit equals 1 divided by one plus RWL, so low-RWL systems must win often and high-RWL systems may win less often and still clear zero.
- A flat round-turn commission raises the required profitability to PC. When average loss is four times commission and average win is twice commission, the floor moves from about 67 percent to about 83 percent.
- The same bound can be applied before a backtest, or when history is missing, by estimating the worst plausible RWL and the commission structure first.
Winners must outweigh losers
A mechanical trading system is a complete, testable procedure for entry, exit, and abstention under stated market and execution constraints. It has a positive net result over a period only when the sum of winning trades exceeds the sum of losing trades.
Expected-value is the requirement that average win times win share exceeds average loss times loss share before and during a position. Profitability is the share of winning trades in the full trade count, expressed as a number between 0 and 1.
The bound without commission
The minimum winning-trade share required for a net profit equals 1 divided by one plus the ratio of average win to average loss. That ratio is RWL.
As RWL rises, the minimum winning-trade share needed for a net profit falls. Low-ratio systems must win often. High-ratio systems may win less often and still clear zero.
Short-term and intraday systems typically show RWL values between 0.25 and 2. True trend-following systems typically sit far above 3.
Holding period changes what history can pin down
In shorter holding periods, exits based on fixed percentages or increments of entry price tend to keep future win-to-loss ratios close to historically observed ones. Staying above the required win share is then enough to keep the system profitable.
In trend-following systems the future average win is unknown because trend size is unpredictable and trailing-stop results depend on volatility. Historical profitability must exceed the minimum implied by historical RWL to leave room for smaller future trends.
Commission lifts the same bound to PC
Commission-analysis accounts for implementation cost so that a required win share is computed after commissions, not only on raw trade averages.
A flat round-turn commission raises the required winning-trade share by adding the commission-to-average-loss ratio to the numerator of the same bound used without cost. The result is PC, the minimum winning-trade share required after that charge is applied to each trade.
When the average loss is four times commission and the average win is twice commission, the required winning-trade share rises from about 67 percent without commission to about 83 percent with commission. When both the average win and the average loss are ten times commission, the required winning-trade share with commission is 55 percent.
Apply the bound before an entry rule
The same win-share bound can be used before a technique is backtested, or when historical data are unavailable, by estimating the worst plausible win-to-loss ratio and the commission structure first.
All readings on this track · 22 readings
- 1995A pre-trade checklist that bounds loss before the order
- 1998Ledger audit of exits, payoff, and overlap
- 1998A return-to-loss filter for drawdown-aware evaluation
- 2000Pair historical volatility with return-to-loss filters
- 2001Credit-spread construction that can fail before any order is sent
- 2002Evaluating mechanical systems in a traders market
- 2002Profitability as a bound implied by RWL and commission
- 2004A day-trading breakeven matrix for size and win rate
- 2006Sit out, size and expectancy as one procedure
- 2006A testable intraday procedure from setup to stand-down
- 2007A planned liquidity offer at the inflection point
- 2011A style-neutral expectancy filter for system evaluation
- 2011Separate buying power from posted risk capital
- 2012Design before you trade: testing mechanical systems
- 2014Ideal trader hindsight as a pretrade filter
- 2014When expectancy and drawdown limits disagree
- 2015Signal, confirm, and invalidate before the trade
- 2015Price the win, stall, and loss before a stock entry
- 2016Construct expectancy by bounding losses and winner size
- 2017Estimate expectancy before you accept the trade
- 2017Size ladder tests for drawdown caps and expected value
- 2017Evaluate a high-yield correlation break as one locked procedure