Profit Factor and Its Merits and Demerits ~ Is it wise to avoid EAs without a stop loss? ~
About Profit Factor
The sample isAMSERdistributed on FX-ON. It is the EA that achieved the top real-time profit and loss in front-end testing for one month.
This timered framewe will look at that part.
Explanations of each item
Bars in test
Number of bars used in the backtest
Ticks modelled
Number of simulated ticks.
In total ticks, it becomes about 50 times the number of 1-minute bars, regardless of the leg.
Modelling quality
Data quality.
Due to the tick generation relation,1-minute bars have a 25% upper limit, and others 90%respectively.
If this value exceeds, ticks are modified.If it is too small, data reliability decreases.
The number of bars where the close of the previous bar differs from the open.
Ideally it should be 0, but in reality it can deviate.
If too many, data reliability decreases.
The equity at the start of the backtest.
If the initial depositis an order of magnitude larger than the funds actually to be operated, beware.
Spread applied to the backtest.
In the current value, it is the spread at the start of the backtest.
In principle, if fixed, adding 4–5 to the fixed value brings the actual performance closer.
Profit or loss after subtracting the initial deposit from the balance. The actual profit.
In this EA,Initial deposit 1,000,000 JPY + approximately 11,810,000 JPY in profit = approximately 12,810,000 JPY in balanceis the result.
Gross profit
Total profit
Gross loss
Total loss
Profit factor
Net profit ÷ absolute value of total loss
When there is profit, it exceeds 1.
Net profit isTotal profit + total loss (negative)respectively.
Net profit can be easily increased by increasing the lot size.
Profit factor calculates a rate, so changing the lot size does not affect it.
From that, the idea arises that a superb EA has a high profit factor.
And creating such an EA becomes a trend.
The method to dramatically raise the profit factor is to reduce stop losses or eliminate them altogether.
However, if you remove stops, profits may not appear → positions accumulate → eventually the account balance falls below required margin and a margin call,
and in the worst case, a margin call may be issued.
- Check if backtest trade history shows a stop loss being set
- Check if the stop loss is not excessively far from the entry price
- See if a stop loss is actually being performed in the backtest
By confirming these, you can verify that an appropriate stop loss is set.
Grid averaging (Nanpin/Shingling) and Martingale mechanisms
Among EAs that reduce stop losses, a commonly used method to reduce failures and increase profits is
grid averaging and Martingale.
Grid averaging is a method of buying more when losses occur to lower the average purchase price.
The doubling entry method doubles the last purchase amount each time.
Entry amounts can grow extremely large, upsetting the balance between required margin and profit.
Monte Carlo methods to reduce entry amounts, or Kokomo methods to increase profits were considered, but all
have the characteristic that profits relative to the initial margin are small and the risk of failure is high.
Also, dollar-cost averaging (more accurately, yen-cost averaging) over regular intervals is also a form of grid averaging. It is suitable for long-term, low-maintenance trading with high-swap currency pairs while managing the spread.
Avoiding grid averaging and Martingale is a wise choice.
However,the really scary thing is an EA without a stop loss. Please check before deploying the EA in real trading.
