trading advisory evaluation

i am interested in getting feedback for an outline of subjects that should be covered in reviews of advisory service free trials. appreciate your time.

max

That link doesn't work for me Science Trader.

quote:

---

Originally posted by day trading

That link doesn't work for me Science Trader.

---

There shouldn't be a dot after "com". I hope this works better:
[url]http://scitra.blogspot.com[/url]

Now I see what the problem was - I hadn't noticed the dot. It's fixed now - thanks.

Your blog covers a number of topics on trading. Could you possibly summarize your Collective2 findings for us?

One of the problems that I had with Collective2 was the "forgiving" of erroneous trades due to mistakes by the operator: Foliage's E-Mini NQ100 trading system. IMHO you can only do this if the operator of the system can prove to you that they would also want possitve results from mistakes reversed out.

I worked for a large hedge fund that ran a quantitative model and if we made a mistake then the clients portfolio was credited if the result was positive but if the mistake yielded a negative result then we took the hit and not the client. This is how mistakes should be treated on systems such as that - in my opinion.

Sure. My experience is that it does a good job at hosting a wide range of trading systems and showing a reasonable set of statistics. It also provides various services to "auto trade" systems, either through an application running on your own PC (TradeBullet) and connecting to your boker (by API); or through their server-based services (which doesn't require you to run something on your own PC).

I think the main problem with Collective2 is that the number of bad systems outweighs the number of good systems. In fact, I wouldn't subscribe to 95% of the systems. But those 5% that do well, are worth trading I think. However, the odds of picking a profitable system at random are pretty bad, so a lot comes down to your own judgment.

Second, C2 does not really have an infrastructure for fast scalping systems. The delay before the signals are executed by an auto trading application can be several seconds (although they claim it's faster when using their server-based auto trading).

In case real-life fills differ from what is shown by C2, it is almost always a scalping system or the result of a vendor trading a market that is not liquid enough to facilitate a substantial number of subscribers (e.g. a system that trades stocks with a daily volume of 50,000 shares). They try to get around that by showing a "realism factor". Also, in case you want to dispute a fill, you can report it.

I also have the impression that many subscribers are quite naive. They expect to find a trading system that will give them a 100% return per year with a Sharpe ratio of 3 for $100/month. That's unlikely of course.

In sum, I think it's a great service, but it really pays off to do a bit of research before subscribing to a system and trade it in a demo account for a few weeks to get a feel for how it does in real-life. Also, every once in a while a profitable system is terminated b/c it got an offer from some hedge fund.

Thanks for that great summary.

Do you ever look at the drawdown and risk of the system. I have noticed that new systems appear on Collective2 and they do fantastically well for a while with phenomenal returns but you often see the drawdowns and risk of these systems at the extreme and very high level. It's my opinion that it's just a matter of time before these systems crash and burn. However, in the meantime, the subscriber numbers grow on this high risk strategy eventually to burn them all.

I think that it would be useful for new investors to see the history of similar systems that have crashed and burned before they plough their savings into these high risk systems. Of course, if there are high risk systems that are doing well for long periods of time then those can be highlighted as well. It seems too easy for a vendor of a burnt out system just to come back with a new name and the same system and start it all over again once their system has burnt out.

You also mentioned that every once in a while a system will be terminated because it's been bought out by a hedge fund. Are those systems and their performances still on there up to the termination date? If so, do you have a link for one? I think that we'd find examples of those sorts of systems very interesting.

Yes, I look at drawdown and risk (e.g. VaR) as well. Some vendors use a high amount of leverage, I guess in an attempt to catch the greedy subscribers. In some cases, even though the resulting drawdowns are large, it can still be worth trading such a system with a smaller amount of leverage.

Indeed there are quite a bit of examples of systems that seem to do extremely well for a while and then suddenly "blow up". In some cases you can see it coming as they make their profits by averaging down aggressively. This will work for a while, until some day they run out of capital while averaging down and have to close for a big loss. Because you can see the individual trade data on the C2 website, it's easy to spot these systems and avoid them.

In other cases it's more difficult, and therefore I personally only consider subscribing to systems if they have a decent track record, e.g. at least a year, preferably longer.

An example of a vendor (a CTA) that supposedly started working for a hedge fund is:
[url]http://www.collective2.com/cgi-perl/systems.mpl?want=publicdetails&systemid=21573498[/url]
(Notice that there are 3 other systems by the same vendor). I think the reason this system was attractive to a hedge fund is that is quite scalable and has a very low drawdown.

Vendors can launch a new system if they crashed, but this new system will be linked through the vendor's name to any previous system(s). There's also a ranking score per vendor that takes into account all his systems. A vendor has to register through a credit card, so his details are known. The only way to hide the link to the old crashed system would be to register with a different vendor name, different credit card, name and address. I have no idea if that happens often.

Very impressive track record Dustin Dubia's system has. I was impressed by the low and normal risk rankings for each trade. I hope he does well.

Thanks for all that info Science Trader!

The extreme-os discussion has been moved here: Collective2: extreme-os