1

Is a scalping EA being tested with ideal spreads, i.e. 0.5 or 1.0 pips? Your broker does NOT guarantee a tightly fixed spread so why are you testing with 0.5 pip spread? Backtesting a scalping EA under ideal conditions will produce false promises and will cause you to lose money in the long run.

2

Does a scalping EA use stop loss less than 6 pips? Don’t get fooled by EAs with smaller stop losses. Your EA may work during back testing but will not be profitable under volatile live market conditions. Tight stop loss does not provide enough breathing room for market pullbacks.

3

Does an EA behave similarly when trading with 1 lot vs. 0.01 lots? Brokers will seduce you and allow an EA to win when trading with a micro lot. When trading with larger mini or standard lots, slippage will occur, causing an EA to fail. This is known as a scalability test.

4

Does the EA demonstrate that it can print money with very low drawdown and no risk? Similarly, if you own a goose that can lay golden eggs, would you sell it? If the results are too good to be true, then it is not likely to be real. You should examine the EA with fixed lot testing to reveal its true form.

5

Has it been tested with more than a 10-year trading period? Don’t get fooled by EAs that are curve fitted over a few years. How does it perform during 2002 Equities Sell-off, 2006 Emerging-Market Crash, and 2008 Subprime Meltdown? History tends to repeat itself, and repetition is the mother of learning.

6

Does it pass the robust test? A robust EA will provide similar profits for Open Prices vs. Control Points model back testing. If you experience large discrepancies, then your EA is broker-dependence, and it may not work in 6 to 12 months. Every year, brokers make changes to their hardware, software, liquidity providers, and etc. It will be a shame to lose all your profits of the past two years because your EA is not robust enough to handle the technology changes.

7

Does an EA require optimization every few months or weeks? This is a sign of instability since optimization should only be done once to a maximum of twice a year. You should optimize your EA when it is recommended by the Walk Forward Optimization.

8

Is the EA promising instant profits within a few days or weeks? Rome was not built in a day; hence, it is important to be patient. Understand that all trading systems will go through some drawdowns and also some stagnation periods (no profits for a few months). Please perform or inquire about these criteria before using an EA.

9

Does the EA have more than 90% winning trades? Don’t get fooled by EAs with no stop loss or large stop loss. One losing trade can wipe out your account and annihilate your months of patience and hard work.

10

Is it being tested with a higher spread to ensure robustness? It is important to test your EA with a higher spread to ensure that it is not a broker-dependent EA.

11

Has it been tested with more than a 10-year period? Don’t get fooled by EAs that are curve fitted over a few years. How does it perform during 2002 Equities Sell-off, 2006 Emerging-Market Crash, and 2008 Subprime Meltdown?

12

Is it still profitable when tested with fixed 0.1 lot? This test will eliminate martingale, grid, and average trading. Don’t get fooled by the promises of easy money and outrageous gain.

13

Does it make more than 500 trades during a 10-year period? With fewer trades, the EA is more likely to be curve fitted, and it does not meet statistically the law of large numbers.

14

Does the largest winner of this EA account for more than 5% of the total profit? Developers tend to select large trends for most of their EAs’ profits. A large trend does not occur all the time, and this is known as selective bias. So avoid EAs with large take profits and large stop loss.

15

Is it providing similar backtesting results for Open Prices, Control Points, and Every tick? If it is only profitable with Every Tick, then this EA is NOT robust. With different brokers, price feeds, internet connections, and etc., this EA will tend to lose money unless everything is under ideal conditions which are not likely to occur with online trading. You must take Murphy’s Law into consideration.

16

Do its backtesting results match up with its real time trading results? You may look at 6-12 months of a profitable trading statement, but the true test is over a long term 10 years backtesting. You must ensure that your backtesting results are valid.

17

Does it trade based on multiple timeframes to capture the different trends of the market volatility? Using multiple timeframes, it will reduce the overall drawdown of your trading results.

18

When tested with fixed lot, does it provide a return/drawdown ratio greater than 25 to 1? You will have an excellent EA if your reward to risk ratio is greater than 30.