Stars Of Collective2 - Issue # 5

I’ve just published issue 5 of My “Stars Of Collective2” web page.

After issue 4 in July we had a long forum discussion (61 posts!) with some really valuable feedback. Thanks to all those who contributed.

As a result, I’ve made a number of changes, including focusing more on recent performance results, to make the list more useful to potential subscribers.

For those who like charts, thumbnail chart images now pop-up when you hover your mouse over the blue-highlighted system names.

Mature systems that are on the best systems list include (in order of ranking):

* [LINKSYSTEM_31243535]

* Trending Futures

* VT26

* Gold Survivor DayTrader

* Turning Points

* TrendSensor Crossbo

* Liquidity Diversified Portfolio Seleukos

* VX-01

* S&P Mini Rainmaker

Congratulations to the owners of all the systems mentioned above, and especially to the Stars Of Collective2 (Trending Futures and Turning Points) - I hope you are all gaining subscribers steadily. Certainly, you are contributing to the reputation of C2, and helping it become the “go to” place for the best trading systems in this challenging market environment.


You’ll find the latest Stars Of Collective2 update at:




Hmmm, your own system is on the list. Did you optimize your ranking criteria around your system or your system around your ranking criteria?


Who cares?

Check previous ranking

Jan 2008

8 systems from the list now:

2 aren’t visible now.

4 failed (1 star of C2 is included )

1 nearly b/e

1(!) is still profitable.

Looks like random choice from C2’s Best List with addition of advertising "TrendSensor whatever"


Murray, can you tell me why Corporate Investments is no longer on the list? I can’t work out by which criteria it no longer qualifies.

Max DD over 30% (34%+ for Corp Inv)

I guess there’s a common thread between what I believe the criteria should be for a good system, and how I design my own systems - so in that sense there is probably some self-selection element.

But I certainly haven’t consciously tailored one to the other. There’s no point really since no-one can subscribe to the system of mine that’s included!

VX-01 is a not a good choice. It had several good months in the beginning, and has meandered for a very long time.

TrendSensor Crossbo is not viewable, so should not be included here.

Trending Futures seems dead.

Seleukos spent too much of its life doing nothing. It had a several month runup, nothing else.


Good point about the Jan-08 list. I guess you’d see the same thing if you looked at what was in the Jan C2 Best Systems list and checked where those systems were at now.

I’m not purporting to be able to tell the future (if I could I wouldn’t be here on C2), but I do think it should be possible to develop a more robust, stable list than the C2 default ones, and that’s the journey I’m on.

I don’t want to get too hung up about the stability of the members of the list, however. That’s a system developers perspective, and subscribers made it very clear that they expect to have to switch between systems as part of their own portfolio management.

It would be good if at least half the systems in one update appeared in the following one. I’m not there yet, but I do think my approach of continually seeking to improve the selection process with each update will deliver a better list than available on C2 - in fact I’ve just taken a snapshot of C2s Best Conservative Systems list (the closest equivalent to mine) so I can do a comparison at the next update.

Thanks for your feedback.

30% is arbitrary. Your judgment system should not hinge on numbers plucked out of the air. "Sorry, it is 34%" is not a good response.

I agree with all your feedback, except re Trending Futures, but time will tell on that one.

I developed a fair way of expressing the “Muzza Score” as a % of the maximum possible, just prior to publishing this update. I think I’m likely to exclude systems with a Muzza Score under 30% at the next update - but I want to do the next updates analysis first, before finalizing the cut-off figure.

My guess is it’s likely to be in the 30-40% range.

At 30%, Seleukos and VX-01 would be excluded.

Already I have an improvement for the next update! That’s what I mean about this being a journey.

Thanks for your feedback.

It’s not arbitrary.

A loss of 30% requires a gain of near 42% just to get back to the starting point.

At 34%, the gain required to get back to the start is over 50%.

My “judgement” is that any trader putting themselves in that position is either taking risks that are too big (and plans to take more to recover that ~50%!) or they simply don’t have a system (eg: I decided to do some experimentation with one of my forex systems to try and resolve a issue with whipsaws).

In either case, they don’t belong on a “best systems list”. For very old systems a larger drawdown (say 3-4 years ago) followed by exemplary risk management may be acceptable, and that’s an approach I’m considering but there’s quite a lot of extra work in that kind of criteria, so the “approaching 50% gain required just to get back where it started” is it for now.

Of course it is arbitrary! Why is 26% not more appropriate than 30%?? For the same reason that 30% is not more appropriate than 34%… Why not use 15% cutoff? Or 22%? What is the point that an average subscriber would bail? You cannot really have answers to this, so you guessed.

I cannot say I know your exact approach, but I think assigning values to the various widgets of your formula and combining them into a final value would be better than picking arbitrary cutoffs. Let the numbers do the lifting, don’t pick what you consider as good values out of thin air.

For heaven sakes, I looked at the link Eu pointed to, You even have your own "Stars Of Collective2" logo label that vendors can proudly display?

Get real.

If you want to give an opinion on systems, fine. But Collective2 has enough mechanisms with the Grid and other "Find a System" logic on the left without someone creating an add-on business out of it.


Your approach would require the calculations to be done to all C2 systems, or at least all profitable ones. Then what? Produce a ladder from 1 to x hundred?

Is a DD of 40% 3 years ago as important as one a week ago? If not how would you do the weighting? I’m open to a change in approach, but you need to be more specific about how it might work.

I have no interest in producing a ladder of hundreds of systems, so an arbitrary cutoff on some final overall score would still be needed. It doesn’t really matter how mechanical you make any process, there’s always a judgment element.

The same goes for robot trading. Folk thinks these just use “facts”, but they all have assumptions and judgments built into them.

I’ve explained my rationale for the 30% DD: it’s the point where I think traders are likely to take larger risks in order to quickly try and regain the ~50% they need to just get back to where they were. Yes, this is just my view of the psychology at play and I might be wrong, but there’s plenty of examples of systems on C2 where vendors have suddenly broken their rules trying disparately to claw back losses quickly. Sometimes they’re even successful, but I don’t think getting lucky warrants their inclusion.

Hopefully astute subscribers will have bailed long before a 30% DD and before a melt-down unfolds out of last-ditch efforts.

As to your other questions: eg why not 26% instead of 30%? At 26% DD the recovery gain is around 35% - still a lot, but not as daunting as approaching 50%.

Here’s some figures:

DD Gain required to recover

20% 25%

30% 43%

40% 67%

Yes, there is some “round number arbitrariness” in choosing 30%, but it’s not entirely lacking rationale, as you seem to be inferring.

If there’s a subscriber out there happy to wear a 40-50%+ DD and let the other numbers “do the lifting”, I’d like to talk to them - I’ve got a system they might love!

I’m not currently offering the logo use to vendors, and none have asked for it.

As I wrote earlier, I think the “find a system” logic can be improved upon.


duh, sorry, i was looking at Macro’s 21% dd. Both systems just seem to miss out, one has dd 31% but perf to dd ratio is above 3:1, the other has only 21% dd but perf to dd ratio is under 3:1, close but no cigar…

sorry for being snappy - bad day yesterday. Your opinions I am sure assist some people in considering C2 systems. At least you are trying

Murray, couldn’t you work backward, instead of guessing forward, to figure out what to look for in systems that might increase the potential for, at least, short-term future success? The idea would be to take regular daily/weekly/monthly/quarterly/etc snapshots of all systems on C2, import them into Excel and try to determine which factors had the best predictive ability … you could then create a new scoring/ranking system based on those studies. You might even use neural networks and fuzzy logic algorithms. This is similar to William O’Neil’s approach in developing CANSLIM.