"I bought a system to trade the Swisss Frank. The number of systems to be sold was limited to 500. In the beginning everything worked wonderful, the more systems were sold the worse it got, and when all 500 were sold, the system was useless."
Bingo. Exactly. There are good trading systems with zero subscriptions
and there are bad ones (or they will GET bad) with many subscribers…
Actually C2 “featured” one of these as a “success” story. The system
went from +$400K to under $100 in less than two months. As the story goes the (OEX) system had a lot of subscribers. What would be of greater value than posting the number of subscribers is something like the “Realism Factor” that weights the volume in the markets traded vs the contracts/shares in the system times the # of subscribers. So if there are 10 subscribers X 100,000 shares recommended and the volume is 50,000 at the trade price the ATV (Actual Trade Value) is ZERO. Moreover, systems with low ATVs cannot be “featured” trading systems.
This, in the long run, would help C2 be more credible. The more credible C2 is the more likely people are to subscribe.
I agree with Sam that the Realism factor should take care of this effect of the number of subscribers. But it would be even better if the Realism factor would include the actual fills that autotraders get, back-reported by Tradebullet to C2. I’m quite sure that Tradebullet reports the fills to C2, because this is written in the log files.
Since C2 knows the fills, it should be possible to compute the average fill and to base the Realism factor on it. Moreover, it should be possible to base the adjusted equity curve directly on the average fills. That would be much better than the current procedure, which, as I understand it, is to compute the optimal equity curve first and then multiply the profits with the Realism factor.
I think that the direct method is better because of this (based on my limited personal experience): when the system’s optimal curve is flat for a while, and the Realism factor is say 50%, then the subscribers will have a loss. So the subscriber’s equity curve will have draw down. The adjusted system’s equity curve however, is just the optimal curve with the profits multiplied by .50 (perhaps a little more complex, but the result is very similar) and will therefore not show the draw down. It will be flat, like the optimal curve (50% of $0 proft is still $0). So in the current procedure the adjusted curve may display the ‘real’ profit correctly, but not the real draw downs.
Coming back to the number of subscribers, this still could be interesting as a kind of popularity vote. Perhaps that is not a good reason to chose a system, but why should C2 restrict the reasoning of subscribers? E.g. if some rich forex-trading woman wants to know the birthdate of a system because she believes in astrology, then she has my blessings, and I don’t think that C2 should try to prevent her to chose in this way. Similarly with the number of subscribers.
BTW Sam, I’ve heard before that result that 90% of the future traders lose. But I assume that you are aware of the possibility that this percentage is distorted by the fact that the purpose of futures is to cover a risk, i.e. a kind of insurance, which makes losses part of the logic.
Jules
"BTW Sam, I’ve heard before that result that 90% of the future traders lose. But I assume that you are aware of the possibility that this percentage is distorted by the fact that the purpose of futures is to cover a risk, i.e. a kind of insurance, which makes losses part of the logic."
Actually (as I assume you are aware) the %'s are worse for the the smaller traders and are based on speculative traders only. Indeed, this very high % of losing traders mostly end up closing their accounts. Hedgers must register as such and are not part of this equation.
Also, why should a private company, vendor, etc. have to disclose income, # of subscribers, etc? Again, to my way of thinking the only relevance is how it effects system performance.
"Actually (as I assume you are aware) the %'s are worse for the the smaller traders and are based on speculative traders only."
No, I was not aware of that. Interesting!
"Also, why should a private company, vendor, etc. have to disclose income, # of subscribers, etc?"
In my view it would be C2 that discloses information, not the vendor himself. There is no moral obligation for C2 to disclose this information, only a wish from some subscribers. As far as the effect on system performance is concerned I agree that there is no reason (provided that the Realism factor improves), but people can still be interested for other reasons.
So let me change the question: Why should the number of subscribers be a secret? If taxes are the problem, well I don’t know how the US tax laws are, but I suppose that the fact that payments are done via C2 assures that the IRS can anyhow easily access these numbers and relate them to the vendor.
Coming back to the number of subscribers, this still could be interesting as a kind of popularity vote. Perhaps that is not a good reason to chose a system, but why should C2 restrict the reasoning of subscribers?
This is not practical because system vendors have no obligation to practice any idea that would invite the attention, say, of the Gestapo (for some it is their spouseor the IRS. Similarly, C2 has no obligation to practice any idea that would invite the attention of those (including the above mentioned entities) who are envious of its success. Most system vendors are not in a popularity contest but in the business of creating a productive system/method with “realistic” profits.
ps: Productiveness is not only a necessary element of a good system/method, it is the good system’s central purpose. But, a systems/methods productiveness is logically interconnected to its other virtues, both in theory and in practice, since these virtues are expressions of rationality. None can be validated in isolation, apart from others; nor can one practice any one of them consistently while defaulting on the others. An objective criteria should uphold not disconnected rules, but an integrated way of existence, every aspect of which entails all the others.
I might as well throw my 2 cents worth into the subject of subscriber numbers. Everyone I feel have made some valid points here. My view is Collective2’s system of recording participation or review of an individual program is sufficient. We can see how many times a trading platform is viewed and the number of people that are following the program.
If we were to make a comparison to the real world, Collective2’s system has more transparency than what one could expect from sites that track CTAs or Hedge Funds. I monitor www.iasg.com which tracks participating CTAs. One doesn’t get the number of subscribers for an individual CTA, but you do get assets under management totals which in a way is a measure of a CTAs effectiveness.
Paul your point about productiveness is right on target. Collective2 provides us a window into each individual program on this site, and much more data than what one could derive from viewing CTA’s performance. Being able to see the individual trades tells me a lot about a program and is more informative than the number of subscribers paying for the service.
If the motivation of those that want subscriber totals is to determine whether there is enough activity to cover the yearly fee for Collective2, then their focus is too narrow. Collective2 is a training ground. Time and costs are probably not going to be covered through subscribers participation.
I monitor www.iasg.com which tracks participating CTAs. One doesn’t get the number of subscribers for an individual CTA, but you do get assets under management totals which in a way is a measure of a CTAs effectiveness.
Thanks for sharing this. Here is another resource: Fund of Hedge Funds Benchmark:
“Fund of Hedge Fund Benchmark” is an equally-weighted average of over 2,700 funds of hedge funds. Benchmark data begins on January 1, 1996.
The concept is to provide investors, money managers and other interested parties a complimentary monthly review of fund performance in the Alternative Asset Center’s (AAC’s) fund-of-hedge-fund database.
It is interesting to see their key data points.
http://www.aa-center.net/products/newsletters/default.htm
http://www.aa-center.net/Mailings/Newsletter/index_200603/
http://www.aa-center.net/donwnload/aac_directory_of_fund_of_hedge_funds_11_page.pdf
ps: I see a Subscription Frequency: The recurring time intervals at which a shareholder can subscribe to shares or units in the Fund, and Redemption Frequency: The recurring time intervals at which a shareholder can redeem shares or units in the Fund. I don’t see any subscriber numbers there anywhere…
"IF we were to make a comparison to the real world, Collective2’s system has more transparency than what one could expect from sites that track CTAs or Hedge Funds."
OK, but those guys trade actual money. C2 systems sometimes
are way under margined or over volumed compared to the margins
requires or markets traded. That’s where the screws need to be
tighter. Can’t let a system buy 500,000 when there were only 50,000
at that price and then “feature” the system’s “track record”.
Agree. MK’s answer to this one is the Realism Factor (RF). But, RF is not being used to compute the final hypothetical real-time returns (final hypothetical real-time productivity). As such, right now, it is a meaningless number. It is not being used in lists like Good longer term Trading Systems in a correct way. MK’s answer to this one is that RF is a work in progress and so it is not yet available for sorting, calculating hypothetical real-time returns etc. Instead, a minimum standard like RF > 0.3 is being used. So, where is the incentive for the system vendors to be honest and reliable when only such a minimum standard is what we have to adhere to?
In my opinion, RF should be used to adjust the final hypothetical real-time returns (return on starting equity, annualized return, total return and also in W:L ratio calculations etc.,), and then used to compute the various lists.
RF is indeed a great advance in trading system analysis but unfortunaltely that great advance is being set back by imposing arbitrary minimum standards. Also the Sharpe Ratio (SR) calculations seems to be much more accurate now on the surface, but I’m not sure it is statistically valid. It is hard for me to believe that Midas Med-Term (Prudent) (O) which used to have a SR of 2.246 previously now has a SR of 21.335, the 3rd highest among all systems.
ps: Jesus’ question “What shall it profit a man if he gain the whole world and lose his own soul?” is admirably exact. The man who “loses his soul” is, in virtue of such a condition, outside the concept of “profit.” Of course, Jesus’ question is not valid if read to imply a dichotomy between the world and the soul. It is instructive only if taken in one meaning: that the integrity of man’s consciousness, its principled harmony with existence, is the precondition of man’s benefitting from any of the splendor the world holds out to him.
This point applies to all human behavior, not only to the issue of honesty. Just as, in epistemology, irrational mental processes detach a conclusion from the realm of cognition; so, in ethics, irrational action detaches a goal from the realm of evaluation. Whenever an object, spiritual or material, is sought or obtained by behavior in conflict with moral principle - whether the behavior involves fraud, improper compromise, the initiation of force, or any other evil - the means employed, by their very nature, clash with reality and thereby deprive the object in that context of any evaluative standing. Once the guidance of principle is dropped, there is no rational method of evaluating
an object whether it is a trading system or any other entity.
" "IF we were to make a comparison to the real world, Collective2’s system has more transparency than what one could expect from sites that track CTAs or Hedge Funds."
OK, but those guys trade actual money."
Here’s an example of a “featured system”.
BTO 400,000 NGRU 0.46 3/22/06 9:51
STC400,000 0.90 3/24/06 14:18
Profit: $174,960 (which is more than 50% of the
systems total historical profit!)
The actual volume at .46 was less than 40,000. So the total
potential profit was $17,496 (likely much less as an actual
buy order for 40,000 much less 400,000 would have pushed
the market higher).
And this is just one trade. If we want apples to apples real
life track records this is THE issue. The # of subscribers
is way down the list and you can always Pmail a vendor
and ask him anyway.
But since the realism factor is apparently not something from which can shortly expect a good solution, a display based on the number of subscribers is relevant.
What about displaying a kind of “survival percentage”: The number of subscribers who are subscribed to the system for more than 3 months, divided by the total number of subscribers. This could give an indication of how (un)happy subscribers are with the system. And the IRS cannot infer the vendor’s income from this.
Of course there is a problem if the number of subscribers is small. I think the percentage should display NA then.
Jules
Jules wrote:
> What about displaying a kind of “survival percentage”:
> The number of subscribers who are subscribed to the
> system for more than 3 months, divided by the total
> number of subscribers.
I think that is an excellent idea! This statistic would be the
most directly tied to subscriber satisfaction of any of the
C2 statistics.
Jules, I like your idea of survival percentage. Then one could at least infer from ‘NA’ that a system has never had any subscribers. MK has in the past expressed his view that simply publishing the number of subscribers to each system would not fit his business plan, even though it would be interesting to many of us subscribers.
"Also, why should a private company, vendor, etc. have to disclose income, # of subscribers, etc? Again, to my way of thinking the only relevance is how it effects system performance."
Sam has a good point, where else would a vendor be obliged to publish information about the number of customers they have?
So how about making it an option for the vendor to display the number of subscribers.
one benefit in knowing a survival percentage is that, if one knows which system has the largest survival percentage, I would then subscribe to it and do just the opposite of what it recommends for guaranteed profits! Sorry, that’s the way the market works!
“So how about making it an option for the vendor to display the number of subscribers.”
It IS already an option. Any vendor can freely post that info either in the system description or in the forum section of the system.
Ta da! We just invented hotwater…
This topic keeps coming up in every other month. Although it would be interesting to see the numbers I agree, there are way many more arguments against it, then for it, thus there will be no change.
Again, any vendor who wants to advertise the number of his subscribers is free to do so. Because almost nobody does that (except vendors with zero subscribers) the conclusion is that vendors don’t want to reveal that number. And that is the end of the story…
"It IS already an option. Any vendor can freely post that info either in the system description or in the forum section of the system."
I disagree, that is not the same thing as a system vendor allowing C2 to publish the subscriber stats. In the latter case the number can be trusted since it comes from an unbiased source.
Unfortunately system vendors do on occasion lie or at the very least stretch the trust so I if a vendor says “I have x subscribers” I don’t know that to be true (who knows maybe they are playing semantic games and counting non-C2 subscribers or some other BS)
They also may choose to release that information only sporadically at times that are favorable to them vs. C2 which would make such information (accurately) available at all times if it were enabled.
"But since the realism factor is apparently not something from which can shortly expect a good solution, a display based on the number of subscribers is relevant."
This could be solved overnight. If a trade could not have occured
(margin, volume, etc) it doesn’t go into the track record. Or at least
systems with RF’s less than X are flaged and pulled from "best"
lists if the trades were impossible. The point here is the systems
should be ranked apples to apples. If one system is making 100%
per year on trades that could be made, and another is making
100% on fictional trades…well it’s like saying I batted .700 in
little league, so I’m a better hitter than Ted Williams.
CTA’s and hedge funds make actual trades, so the “RF” is built in.
Also, I’ve noticed in this thread there are some great systems with realistic track records that claim ZERO subscribers. What does that
tell us about the value of the # of subscribers? Just because people
want to see it doesn’t mean it’s a good thing. How many Duece Bigilo
movies have there been?
One could extend your contrarian-order logic to any system which had sufficient subscribers to cover the volume you (and other contrarians) wanted to trade. It is not clear to me that simply trading against others who are following the system signals would lead to profits. Perhaps you can give an example in a hypothetical system that uses limit orders to trade a single futures index.
"This could be solved overnight."
But these things were discussed here more than once, and yet the RF still has some serious limitations. Apparantly it is more difficult than we think, or it has no priority, so probably it won’t happen overnight.
"Also, I’ve noticed in this thread there are some great systems with realistic track records that claim ZERO subscribers."
I believe that was just a joke.
