Automated Trading Systems That Work...


I have been trading automated systems professionally for more than one year. During this time, I have purchased and traded more than 20 automated strategies and have found only 3 to withstand forward testing.

What has been your experience? How robust have you found automated trading strategies to be? Do you have any strategies that work better than those that I have?

Thanks,
Originally posted by trading truth

Tomgilb

You have stated that you "used the default settings in the scripts" yet when I looked at your results for the Russell Viper for instance I noticed that on line "p25", "trailstoptrigticks2" you have input the value of "20" when it should be a value of "40" ( 40 is the default setting). This makes a huge difference in the performance of the overall strategy since the value of 20 which you are using would cut short many of the profitable trades that the Russell Viper strategy goes on to make by utilizing the trailing stop feature. I submit to you that this is partially the reason that you are only showing net profits of approximately $7,000 for the period from October 28 to the end date of your data, when in fact the strategy results show net profits of approximately $18,000 for that same period.

You also stated that "I did not exclude the news times on any of the scripts because on average you win sometimes and you lose sometimes". Unfortunately, this is not how the scripts are to be run, and the instructions given in the printed material and the webinars are quite clear regarding the other strategies that adherrance to the time settings when the strategies are to be traded and not traded is of paramount importance to the peformance of the strategies. Your decision to ignore the news times when the various strategies should or should not be traded have totally made the results you have posted not only totally incorrect and meaningless, it again reiterates my previous statement that people should not be posting results of strategy back tests on this forum or any other forum for that matter if they do not follow the recommended trade times and settings of the strategies they are testing and then go on to make claims that the strategies that they "have tested" do not match the published results of the strategy provider.

As a final note, I wish to emphasis that the results that I previously posted regarding the performance of the Russell Viper system ( up approximately $18,000 since late October) are the result of a combination of both my own personal back testing on a day be day basis, and real time live trading using the default settings in the strategy and following the instructions for its use to the letter. My results for this period closely match the results posted by the strategy vendor, within a few ticks on a day by day basis.

Your choice to subscibe or not to subscribe based upon upon your flawed back test results obviously is your personal decision, but it is unfortunate that you have attempted to influence the thinking and decision making process of others who read the posts in this formum by posting results which are inaccurate because you did not follow the rules and instructions for the strategies that you "backtested".

It should be noted that since the Russell Viper strategy was released, it is up over $9,000 which is over 30 times what the monthly subscripion fee is for this strategy and which also includes all of the other 5 strategies which are also included in the subscription price.

Res Ipsa Loquitor ( "The thing speaks for itself")


trading truth,

If the default trailstoptrigticks2 is 40 then that is what I used because I did not make any changes to the default parameters. If the default is 40 then the value of 20 in cell p25 is a typo.

If you win sometimes during news and lose sometimes during news, there would not be a significant difference in the overall performance.

If I trade the RussellsViper on the day prior to FOMC announcements, there would not be a significant difference in the overall performance because FOMC announcements happen infrequently.

If I feel that my results correctly represent the overall performance of your autotraders, despite the vendors' published results, I am free to post my findings on this or any forum (at the moderator's discretion, not yours), as a service to fellow traders. It is up to the reader to decide whether to believe me or you.

If I had actually traded these autotraders, my PnL would exactly match the results I've posted, not necessarily the vendor's published results, because those data are taken from my charts. For me, this is the overarching consideration.

I stand by my posted results as being sufficiently accurate for me to make the proper decision not to subscribe to Viper's autotraders.

-Tom

The big money is made by selling subscriptions to a trading service and not risking your own capital.

My intelligence sources say there are approximately 250 paying subscribers to this service. Do the math at $299 per month.
Originally posted by phantasmagoria

...approximately 250 paying subscribers to this service. Do the math at $299 per month.

$74,750/month
$897,000/year


Kel,

Are you saying you are willing to accept a $12k drawdown on a $15k acct? Wow! Too rich for me! If you started at the wrong time you would only have $3k in your account. That is a blown account by my standards.

My experience with autotraders so far is that they all have big drawdowns and/or wide stops. This is because the market is too dynamic to fit into an algorithm. The concept of an autotrader is intoxicating because we all want a money tree. But controlling greed is a key ingredient to successful discretionary trading, and it should also be a personal goal in all areas of life, IMHO.

-Tom


Tom - I began trading the above system based on a historical worst case drawdown of about $6000 (based on trading 1 contract), I then multiplied that by 1.5 times and added margin requirement and that was my minimum starting account equity - if I exceeded the 1.5 times max historical drawdown on a 1 contract basis I deemed that my system no longer worked and I stopped trading it - based this off of guidelines from one of Larry Williams books...the results posted above are the so called walk forward results...after the system was developed...I followed it on paper before trading it with actual money but eventually I stopped in first quarter 2008 as I exceeded the money management drawdown I specified in my trading plan.

With regard to picking a wrong time to start trading you never really know but if you wait for a $3000 drawdown on paper before following the system with real money and your trading plan says you will stop trading after a $6000 drawdown on a single contract basis and you system immediately blows up reaching that $6000 drawdown you have saved yourself $3000.

Knowing that risk ahead of time is valuable info and the only way I would trade a mechanical system. And you're right mechanical systems seem to have high drawdowns eventually; even when historical testing show promising results. As an interesting excercise take a look at www.futurestruth.com real time tracking of the publically available trading systems and the track record of some of those systems - even the higher ranked systems live through some pretty extreme drawdowns.

In the past the only way I traded was as a automated or mechanical system trader - I now find myself trading on a discrestionary basis and doing much better...

However that said there seem to be some incredibly talented and successful traders contributing to this forum.

Best,
Kel
Kel,

I looked at futurestruth.com and you are right, even the better ones have big drawdowns. But at least one of the better ones, tradingvisions.com, provides extensive professional past performance tables and calculations, something Viper Trading Systems' eminiautotrader.com replaces with only cheesey charts just a few weeks back. What a difference, and the cost is half.

So I am not finding any autotraders that are sophisticated enough to avoid losses in chop, that find the trend, and that are consistent with the equity curve over time. I'm not sure it's even possible.

I've even built a few simplistic autotraders that seem to perform as well as many out there you can pay for, but still none are up to what I want. Discretionary trading still beats mechanical autotrading.

Your loss Tom, that system popped another 4K+ last week. Well worth my $299. Maybe I got your spot!!
Originally posted by Butthead


Your loss Tom, that system popped another 4K+ last week. Well worth my $299. Maybe I got your spot!!


Based on long term historical trades,(of which they provide none) there will be huge drawdowns eventually. Be careful.
Originally posted by tomgilb

Kel,

I looked at futurestruth.com and you are right, even the better ones have big drawdowns. But at least one of the better ones, tradingvisions.com, provides extensive professional past performance tables and calculations, something Viper Trading Systems' eminiautotrader.com replaces with only cheesey charts just a few weeks back. What a difference, and the cost is half.

So I am not finding any autotraders that are sophisticated enough to avoid losses in chop, that find the trend, and that are consistent with the equity curve over time. I'm not sure it's even possible.

I've even built a few simplistic autotraders that seem to perform as well as many out there you can pay for, but still none are up to what I want. Discretionary trading still beats mechanical autotrading.

Tom, have you searched Collective2?
Originally posted by phantasmagoria

Tom, have you searched Collective2?


I found that site some time ago, but I haven't been back in a long while. Thanks for the reminder.

-Tom
Originally posted by tomgilb

Originally posted by phantasmagoria

Tom, have you searched Collective2?


I found that site some time ago, but I haven't been back in a long while. Thanks for the reminder.

-Tom

I lost faith in Collective2 when I discovered that the system owners could "appeal" bad trades and have them removed from their accounts. E.g. if they made a mistake.

The reason that this doesn't work is because if they make a mistake that pans out favorably then they'll never request to have that trade removed from the record.

All mistakes should be included because at the end of the day you'll make the same number of mistakes that work out favorably and unfavorably.
I've been trading for about 12 years and haven't seen an automated strategy that works consistently. The ones that have worked in the past or are making money right now will probably fail miserably in the next couple weeks or months at best. I'd rather rely on my own ability to trade than a computer.