Yee Haw! That thread was getting to be a major PIA just to load in my browser.
Oh Boy!!! Here we go again!
That thread was the only internet soap opera there is!
This time I agree with you 
Sorry, one last comment.
“You ask:
Question. Can anyone with a losing system switch it
to a “test” designation at any time {snip}.
No, a vendor cannot do this. This is why C2 is so strict about test system designation. In order to create a “test system” you need to designate it as such before you issue any trades.
MK"
OK, I was confused. I see why now. A vendor can change the name
of a system at any time, correct?
I just changed “SCook” to “SCook1 Test”.
Did “Midas Med-Term (Conservative) (O)” recently change to
"Midas Med-Term (Conservative) Test (O)”? Or is my
memory failing?
Also, is there a way to be sure the vendors comments align with
reality?
Please see vendor comment: "Sharpe Ratio: 3.659 (integrity)"
C2 stats: "Sharpe Ratio 1.774"
Some might find this misleading. I can see why it could be nearly
impossible to police vendor commentary. Perhaps it should be
policy to NOT allow “stats” in commentary.
Sam
Pal is doing this all the time. He create systems, over leverage them and hope to get lucky. When the system crash, he change the description to include Test in the name and pretend he is just testing things. You will notice he charge $301 for his “test” systems. Not a bad deal to charge subscribers while you are testing your system.
Chris
The way C2 is set up, it is too tempting for some people to load up or swing for the fences in hope of being noticed for subscriptions. In the process some good systems might be lost, but on the other hand the people who do this are most likey not the people to trust ones money on. And perhaps this is one reason why things are set up the way they are, to weed out the venders with possible trading issues or trading personality flaws.
Nothing personal ment here Pal, just reality, something you seem interested in.
Not only can you change the name after the fact but you can stop supporting a system which for all intents and purposes hides a system from potential subscribers.
Pal seems to be familiar with this concept as well but he is not the only vendor at C2 who has done this.
I could be wrong Pete, but you could be referring to the fact that I just killed my test system. This system was created before the option of hidden test systems. Since I now have the test system hidden from veiw I did not renew, but killed it instead. As bad a system as it would have been it was still ended above the S&P and in profit.
Yes you are right however.
Actually I wasn’t referring to you and didn’t know that was the case but thank you for being so forthcoming.
Someone else point out last week a vendor who had stopped supporting a system and then started a new system with virtually the same name - a practice that I find particularlly offensive.
It’s not too difficult to spot something from Pal/Paland… it is typically devoid of meaningful content, with fluff and filler as the main ingredient. Said fluff often takes the form of quotes from famous people taken out of context, and opinion unsupported by facts (other than the mis-quotes).
Should you question him on any of this, the reply will typically be a philosophical answer… again devoid of meaningful content, along with some inane statement demonstrating his inability to differentiate fact from opinion.
I may have mentioned this somewhere before (can’t remember):
C2 keeps track of system name changes. As of today, that data is not accessible to users. But I intend to make it easily available shortly. I suspect you will find an interesting correlation between system names being changed to “XXX Test” and an immediately-preceding equity drawdown.
In any case, I want to use this post as a chance to remind C2 traders that you should not necessarily take a vendor’s word for it that a system is a “test” system. This particularly applies to any system that has been created in the past few months, after the formal “Designate as a test system” feature was made available. (Of course, this is not always the case. The “Designate test system” feature is relatively new, and there are vendors who create publicly-viewable test systems in good faith.)
Remember, if a system vendor truly intends for a system to be a “test” system, he must so designate it before he starts trading with it. If he does so, the system will not appear to the public, and will not affect the vendor’s aggregate stats.
So, sophisticated C2 traders should keep in mind that if you see a “test” system, in all likelihood it was so designated after the system lost money. A real test system will not be visible.
This will become even more obvious once I make the “system name change history” available to users looking at systems.
MK
Speaking of test systems, I did a search based on Sharpe Ratio
and see:
Test SystemSharpe Ratio 1025.94
I understand “THIS SYSTEM ISNT TO BE TAKEN SERIOUSLY”,
but I have trouble understanding how it has a Sharpe Ratio of
1025.94 with a DD of 67% in several days. Maybe this is because
the Sharpe is only updated less frequently, but there are other
systems with losing records, huge DD’s, or other issues that have
HUGE Sharpes. Just wondering if you have taken a good look at
your Sharpe formula.
Thanks.
"…but I have trouble understanding how it has a Sharpe Ratio of
1025.94 with a DD of 67% in several days"
See the thread “Sharpe Ratio Calculation is WAYYYY off!”. The Sharpe Ratios are now computed on monthly periods. The DDs are computed on a daily basis. That’s why they don’t necessarily agree: If a DD occurs and recovers within one of the monthly periods, it won’t affect the SR. This might explain the problem.
There has also been another discussion about the SR: A large profit can lead to a decrease of the SR. So a large profit followed by a loss of the ‘right’ amount (within a month) may effectively cause an increase of the SR.
The fact that the SR is based on monthly periods, together with the fact that it is reported after 90 days, allows the possibility that the SR is based on a sample of only 3 time points. I wouldn’t be surprised if the SR will be quite unreliable in such cases.
Jules
"The DDs are computed on a daily basis. That’s why they don’t necessarily agree: If a DD occurs and recovers within one of the monthly periods, it won’t affect the SR. This might explain the problem."
I don’t think so. Please see:
http://www.miapavia.it/homes/ik2hlb/sr.htm
Sharpe = Excess_return / Annualized_standard_deviation_of_returns
On C2: System Details: Test System
Sharpe Ratio 1025.94
Max Drawdown 61.75% (20060501 to 20060516)
There are prior, multiple, 50%+ DD’s and runups. There is no way
a system this volatile would have a Sharpe of 1025.
Here’s a reality check:
http://www.cta.visionlp.com/pdf/TRS/tradingsolutions.pdf
"Trading Solutions Sharpe Ratio was also the best. It
doubled that of most other top CTAs listed, coming in at an
impressive 2.21!"
Anyway, there are losing or near losing systems, with erratic equity curves, with double digit Sharpes on C2. Sorry, no way that’s right.
My thought, and my wish, is to be able look at the “all systems” page and try to rank systems in terms of risk/reward. The Sharpe Ratio would be of help except "Sharpe Ratio Calculation is WAYYYY off!"
How about Profit Factor (Gross Profits / Gross Loss) rankings? Or better yet, Total Net Profit / Intraday Drawdown? I know Pal and some of the other guys that like massive leverage will say this is “unfair” or invalid, but as it stands there is no way to rank systems by Reward / Risk and
THAT is what it is all about.
Thanks!
Somehow I knew that you would disagree… You clearly didn’t read the thread to which I referred, or you refuse to understand it. The webpage to which you refer offers nothing news to what has already been discussed in that thread.
Jules
I read some of the thread:
Pal suggested a losing system could have a positive
Sharpe, etc. I guess that’s when I tuned out and decided
it might not be worth my time to read every word.
Are you saying, as a PHD is statistics, that the Sharpe
on C2 is correct?
Or are you just being disagreeable with my other suggestions
about Risk/Reward? Understood, some of your pals on C2
agree: Risk is a four letter word that should not be mentioned
in public.
"Are you saying, as a PHD is statistics, that the Sharpe
on C2 is correct?"
No. What I say is that a large DD together with a large SR can happen because DDs are computed on basis of the daily account values and SRs in C2 are based on the monthly account values. So the example that you gave can be a consequence of this choice and does not necessarily imply that the computations are ‘wrong’. I do not know whether that is actually the case in your example. However, like you, I do not like this ‘contradiction’ and therefore I suggested that the SRs should be computed on a daily basis too. There were several people against that; see that other thread.
Jules
"No. What I say is that a large DD together with a large SR can happen because DDs are computed on basis of the daily account values and SRs in C2 are based on the monthly account values."
Again please look at “Test System”. There are 4 closing DD’s of 20% or more and several runups of 200% or more. The Standard Deviation (daily, monthly, whatever) is HUGE on this system. The equity “curve” is one 180 turn after another. Industry standards show the smoothest and best equity curves rarely achieve a Sharpe of 3.0. The “Test System” has a Sharpe of 1000+. Doesn’t this make you wonder?
Bob Fulks (MIT prof and long time Omega systems writing guru) said:
“Sharpe Ratios above 3.0 are outstanding.”
"A trading system with a smoothly increasing equity curve will have very
consistent monthly returns, a low standard deviation of returns, and a high Sharpe Ratio. This greatly reduces the chances of going broke. Smooth equity curves are good. Choppy equity curves are risky. So we should always optimize our trading system for the highest Sharpe Ratio."
So here is the bigger question: shouldn’t C2 try to bring performance numbers into line with industry standards? In the long run wouldn’t C2 be more credible if performance was believable? Wouldn’t there be more long term and happy subscriber/customers if what you saw was what you got?
Another quick look at “Test System”. It trades 200 lots of the DJM6. It has a RF of “76”. A quick look at recent DJM6 charts shows it can take over an HOUR to get volume of 200 contracts. The DJM6 can move over 100 points in an hour. IOW, “slippage” on a market order of 200 DJM6 can easily be 100 points x 200 contracts x $10 point or $200,000.
If anyone thinks slippage on market orders of 200 lots “balances out” I’ve got some bridges outside NY you also want to invest in.
So two requests/ideas:
1) Tighten up the RF to make over-leveraged fantasy trades pay a realistic price.
2) If we can’t see something like an industry standard in the Sharpe then let’s see something very easy to compute like Net Profit / worst DD ratio in the “show all” ranking table. Both Net Profit and worst DD are already computed.
The current setup favors fantasy systems that either aren’t really profitable to begin with or over leveraged systems that are ready to
blowup. There is no good way to find good risk/reward systems in a
C2 search, while garbage systems pop up right near the top all the
time.
Basically I agree with everything you said. The SR of 1000+ does not properly reflect the volatility that one will experience with that system. And this is bad. What I said is that, if one wants to use SR, then the only solution to this is probably to base it on daily periods instead of monthly periods; but some people were against that. Using another coefficient could be an even better solution. This has been suggested in several other threads too.
Jules
