Probacktest.Are we being conned?????

[parbynat]

HI all,
I recently came across the probacktest module on my account with IG index and thought what a nice tool to play around with.So over the weekend i thought i could come up with the best system ever,think we all start off that way.Anyway based on around 4 conditions i came up with a system that gave great results on the ftse100 1 minute chart around 89% winning trades over the last 2 days.So today i was all set up on my demo account to give it a go.First trade 8 points profit,nice start,second trade major spikes down 15 point stop activated,ok no probs.It was at this point i lost my connection.When i got back in and started up my system i was baffled to see that my first trade was showing but my losing was nowhere to be seen.
This has sorta got me thinking about my system but more importantly the validity of the module.
Has anyone else come across this.Would love to know whats going on.
Many Thanks..........

 

[NetTecture]

Given technical problems you indicate - I would not necessarily say that it is a scam. I currently also work in trading system development - architecture wise, writing a framework, and the question how to deal with "blown transactions" is a good one.

I think it is technically valid that the system tries to close the deal and then ignores it. Trade system performance evaluation should ignore "outside factors", and a bad computer crash etc. is definitely something that has no direct relation to the exchanges. It may be the trades get only entered once the trade is closed, and the lost connection may have stopped that.

I would NOT call tha conn without further evidence (like the system killing your network to not have to enter the trade). Rather - get the system moved to a proper location where you ahve a good connection MY trading server will run in a proper data center For exactly this reason - I do not want that crap (i.e. connection problems) on my desk.

 

[£10kLoser]

Back testing in general is highly over-rated. There are bound to be patterns in even totally random charts and to assume that they will continue is just gambling. Most systems devised from back testing are simply 'curve fitted' to show a desired outcome and have no basis in real market behaviour. Even if there was a basis, to assume that the markets will behave the same in the future as in the past is clearly a fallacy.

When you think how many times people have done back testing to produce '89% winning trades' yet there is still no proven system that guarantees profits, you'd think back testing would be less popular as a way to create a trading system.

 

[TheHoneyMonster]

Backtesting is a waste of time.
Trading can only be done successfully by staying in the present moment.

 

[Phil Mibbutz]

Backtesting is a waste of time.

Yes, but at least it's not a waste of money...

 

[NetTecture]

£10kLoser, you totally miss an important point in backtesting:

* You can check how your system works in various conditions. Granted, most backtesting tools do not support that. This is why I write my own. Then I can check exactly how a system works in defined conditions. Not to curve fit, but if it only works in trending, to start with, I would like to find out BEFORE LOOSING MONEY.

* You can kill those systems that loose money. Without forward testing them a year or two to be sure they work.