Prepare to say goodbye to demo testing

AsstModerator

FPA Forums and Reviews Admin
Messages
5,746
I've been waiting a long time for this announcement.

The FPA exists to protect traders. As problems have come up and abuses have occurred, the FPA has had to tighten the rules for Performance Testing.

When the FPA Performance Testing program began back in 2009, getting volunteers was hard. The FPA had to buy a lot of the EAs. Volunteers with live accounts were rare, so demos were permitted. When Signals and Managed Accounts were added, those also permitted demo accounts, since it was the only way to get volunteers.

Later, the quantity and quality of volunteers improved. More live accounts were submitted. Managed Accounts were taken 100% to live accounts after a few incidents where client accounts were traded differently from the demo accounts on display at the FPA. Later on, Signals were also moved exclusively to live accounts after some issues.

New rules banning demo restarts had to be implemented. This was because some companies ran high risk trading schemes and changed the investor passwords the moment the accounts began to lose money. Other companies crashed account after account and always wanted a new one on display as quickly as possible.

Last year, new limits were introduced on demo accounts. Submitted demo accounts could be no more than 2 weeks old. This was because a lot of older accounts came in and immediately stopped performing as well as their history showed. It was plain that some of these were the the best performers selected from a large group of accounts all running different settings. All demos were given a maximum run time of 15-16 weeks. An exception was made for demos running in parallel with live accounts.

Recently, one company was caught trying to slip a demo in under a different domain name. They claimed to be unrelated, but there was more than enough evidence to link them. Another company showed large profits on the first 3 trades, all of which were far larger than the ones taken by the EA. In retrospect, it was obvious that they started 8 demo accounts, had half go long and half go short on a large trade. The 4 that won were split short and long on a second trade. Then the last winning pair was split on a 3rd trade. The account that won all 3 trades had an artificial profit spike to make it look successful before the EA was installed. There have been several scalping bot demo tests recently that had excellent demo tests and reviews complaining that they didn't work on live accounts.

There are too many bad companies out there finding new ways to abuse the system. The time has come to tighten the PT rules again.

The FPA had a meeting last week. One of the main items I proposed was phasing out demos. My plan has been approved.

As of June 1st, 2013, the FPA will no longer accept any new demo accounts for performance tests. Existing time-limited demo tests will be allowed to run for 15-16 weeks and then will be shut down. On or before September 30th, 2013, all parallel demo tests will be stopped. At that point, the FPA's Performance Testing program will be 100% on live accounts.
 
Excellent decision FPA...I back this 100%. In my experience with EA development - a live environment is so different from a demo one that it almost makes testing on a demo account redundant. How a live broker's environment responds to new events, spikes, market opens, spreads, execution response times etc all makes a huge difference as to how an EA performs overall ...so much so that I only really use a demo account for development and final system testing. There is no better way to test an EA than to put it up on on a live account with low lots and manage the risk accordingly. The chance of losing real money certainly focuses one's attention on the finer aspects of an EA's operation.
 
That's the sensible decision since Demo accounts are no different with "Play money" in poker where one will tend to go "all in" more often knowing that one can always request for more "play money" when the account is wiped out.

That post on "USD10k to USD10m in 10 days" is one very good example. Try that on a real account and the probability of account wipe out is more than likely.
 
Back
Top