I’d like to bring up the issue of rescaled futures systems on C2. When a futures system is rescaled by the manager the historical positions are changed to be fractional futures (if a single future or odd lot was traded) and the drawdown is reduced by the rescaled amount. The issue is that the subscriber is given the impression that the max drawdown was MUCH lower in dollar terms than it actually was in real life and lower than it could be even theoretically since fractional futures don’t exist. For example, if I’m trading single lot ES futures on a 100k account and regularly hit 50% drawdowns ($50,000) I can simply rescale my system down by 50% and my past drawdowns now look like they’re only $25,000 even though the system is still trading single lot futures and the risk is now double my past drawdowns in percentage terms, or 100% of the account is at risk. This is very deceptive in my opinion. Take a look at Maximus Decimus as an example of this (https://collective2.com/system93232156). If he hadn’t rescaled subs would have seen the real risk of a $80k drawdown in the past instead of only a $20k artificially rescaled drawdown. Since there is no way for the subscriber to trade at less than 100% of the strategy since fractional futures don’t exist showing a $20k historical drawdown was not realistic in any way.
As long as the % doesn’t change, I don’t see your problem. Whenever I look at a system, I look at %s and not absolute numbers. Absolute numbers mean nothing since we are talking about sims.
When there is a system with 100K account trading 15-20 cars, I always calculate the per day per contract number and that gives me an idea just how much I would make following it with a small account.
My only problem is with rescaling that the subscription fee changes the % and by quite a lot. If it was 1-3%, I wouldn’t mind but apparently the change can be 1/3rd of the results. It also changes the DDs…I just don’t see how a 100-300$ fee changes the result so badly…
You completely missed the point. Absolute numbers mean everything when you’re trading futures unless you’re trading with millions of dollars, because the minimum number of contracts, and most frequent position size of contracts traded is 1, not .025 contracts or 10% of a contract. With stocks, options and forex its a different story because you can scale the position size down to a very small size without changing the drawdown percentage. For example, I can trade US Stock trader https://collective2.com/system64928903 , on a small account by scaling down to 1% when I subscribe because he’s trading 272 shares and the drawdown percentage expectation will remain the same even at 1% trading size. With ES futures, each single contract is worth about $100k and it’s impossible to scale down less than 100% in real life. When a manager “rescales” an account and create fractional futures then you are left with the impression that you could have traded the account at a smaller size and had a smaller absolute drawdown than was possible. The absolute drawdown amount that you could ever expect is MUCH larger than the performance history shows. This is only true for futures.
It is completely irrelevant how much a futures contract is worth. What relevant is the margin requirement at your broker so how many points you can afford to lose before you have to stop trading. Since you don’t understand this simple principle, I see the problem of you not being able to work with %…
If a vendor has a 100K account with 40% DD and you subscribe to it with your 25K account, then you can expect a -10K DD. I don’t see what the problem is. If the vendor rescales and the account is now 50K, you still can expect a 40% DD, the % didn’t change…You just apply that % to your account size…
When a manager rescales the account, just wait a few days or until the new trades show up and you will see how you could trade the system. I rescaled in October, now if you look at the October trades, you can see if you could use those trades in your account. Yes the September contract numbers aren’t full numbers, but unless you want to calculate per contract return, it doesn’t matter. Since I am only using 1-4 contracts, you can only upscale your leverage, for example…
“If a vendor has a 100K account with 40% DD and you subscribe to it with
your 25K account, then you can expect a -10K DD. I don’t see what the
problem is. If the vendor rescales and the account is now 50K, you still
can expect a 40% DD, the % didn’t change…You just apply that % to
your account size…”
Not exactly. If a vendor resized to 50K, the expected drawdown in your example is now 80% not 40%, due to the nature of trading futures which cannot be resized in real life lower than 1. So in effect, his position size has doubled from the historical performance (assuming single futures were traded) but with no indication of this happening. To make matters worse, there is no way of telling by much the vendor rescaled. Was he trading at $1m in the past or $100k in the past? Without this information the past performance is not meaningful in terms of drawdown and expected volatility.
You’ve got it backwards. Margin requirement is irrelevant to the drawdown you will experience. It is the underlying contract value that determines the volatility and drawdown.
Holy molly! AphaTradin, it is your account value and the current mark to market of your position that determines the size of your DD, not the underlying contract value. Seriously people, where do you get these ideas? Please tell me you are a high school dropout…
OP, I am telling you from both experience (I did rescale) and from 3rd grade math class knowledge that rescaling itself doesn’t change the DD%, end of story.
I rescaled and my 25% DD stayed the same. Do you believe me or do you believe in your misconception? Anyway, if you don’t get it, there is nothing I can do to explain it better…
Yes but in your original post you said it was the margin requirement was the most relevant. Margin has nothing to do with drawdown. Re-read your post.
Your history is now pure fiction: the $3000 drawdown in September never occurred. It was $6000 at least. And it never could have been $3000 because fractional futures are fictional. So the drawdown is misleading. Anyone who didn’t know about rescaling would assume your max drawdown was $3000 and they might size their account accordingly which would be a huge mistake since future drawdowns are much larger. That’s the point you’re missing.
It seems like that once a system rescales the dd statistical figures, both in absolute and percentage terms, become meaningless. Personally, I would prefer that C2 does not allow rescaling. I also prefer futures system which do not increase the position size as the account grows but let the subscriber decide if and when to scale up. In that case it would be up to the vendor to periodically post what the return would have been had he scaled up at certain intervals.
Pedro, if the sp500 stock market drops 3% the @es will drop $3000 and the @sp will drop $15000 because the underlying contract is different by a factor of 5. That is why the underlying contract is what affects your dollar drawdown and volatility. The margin requirement is irrelevant to the $ loss you will experience. I hope this makes sense now. And btw, I have passed the series 3 licensing exam for CTAs.
Yes I agree it shouldn’t be allowed for futures systems. It is extremely misleading. If you look at some of the past leading systems like Maximus decimus you can see how deceptive past drawdown are when rescaled.
Now that explains things.
BTW, recently you pointed out deceptive practices from new vendors by gambling in the beginning and then when it turns out well open the system to subscribers. Take a look at the previous forum post, Draken ETF. He decries this practice too, “While studying other systems currently on C2, I realized many equity system creators were using tricks to enhance their statistical returns, and entice unsuspecting subscribers”, but then he does exactly that, taking huge risks on his first 3 trades and after it turned out well (after quite a while) he promotes it to the public.
Anyway, keep up the good work by pointing out suspect systems.
Instead of teaching you guys 3rd grade math, I will just quote the owner of this site:
“A historical max-drawdown of 15% remains exactly that… a 15% historical drawdown.”
Are we done yet?
Afraid not. I realize this could be detrimental for you in fooling subscribers and therefore may be contrary to your financial goals but do you think the 23 autotraders of Maximus Decimus wanted to lose $65k in a single month on a system that displayed the max loss as historically only $20k even though it was at least 2 to 3 times that when trading real futures of a single lot size (the same as current)?
Feel free to follow along here:
I will just point out another logical/investing fallacy in your OP, and I am done here:
You think that historical DD is indicative of anything. Just like the “past performance” disclaimer is there for a reason, historical DD isn’t indicative of future DDs. So just because a system has a max. 20% DD in the past that doesn’t mean the vendor can’t have a 50% DD in the future.
This single fact makes your OP moot.
Oh yeah since while I am at it, let me point out the other obvious mistake in your complaint, this supposed “problem” exists for any system, not just futures.So if you complain, complain about all systems…
As I have shown, you don’t understand basic terminology of investments and math. This is where our conversation ends…
Of course past drawdowns don’t guarantee that larger ones won’t occur in the future. But that’s not the point here. The complaint is that much larger past draw downs were hidden from view so subscribers could not make their own judgement about the system’s volatility and likelihood of hitting that large drawdown again.
As far as stocks being the same as futures, it appears you still didn’t follow the other thread about the difference so here it is again:
Ash -
I think we’re talking past each other, and we’ll have to agree to disagree.
If a may summarize the disagreement:
You are saying: “After rescaling, past stats are meaningless, because it is not possible to trade fractional portions of a futures contract, which is what appears in historical data after a strategy rescales its nominal Model Account size. So, after rescaling, historical results are essentially invalid.”
I am saying: “I agree with you, Ash. It is not possible to trade a fractional futures contract. But I am not suggestion that you can do so. What I am suggesting is that one can look at past results at any given point in history, and can measure certain important statistics (such as drawdown) on a percentage-basis relative to the capital that was deployed at that historical point in time. This allows us to make judgements about what kind of drawdowns we might expect in the future.”
Just to take an made-up example: Imagine a system that starts with $100,000, trades for a year, and then rescales its Model Account to $10,000. We may look at its very early trading history and see a drawdown of 20% on the capital used by the strategy. This does not suggest you could have traded the strategy with only $10,000 back then. It only suggests that, given whatever amount of capital was required to be deployed at the historical date in question, a 20% drawdown happened.
I know I’m not going to convince you to change your mind about this, and that’s okay. You feel strongly that Model Account rescaling invalidates past trading results. With equal conviction, I disagree with that opinion, but so be it. I think the best place to leave this discussion is for us both to recognize that we each have our opinions, and that is that. Since I manage this site, though, and since I think there are valid reasons for supporting nominal rescaling of Model Account size – and indeed think the practice can benefit subscribers by allowing them to AutoTrade a strategy using 100% Scaling – I will allow the practice to continue.
I appreciate extremely that C2 is important to you – important enough that you are willing to spend your valuable time offering well-reasoned arguments here on the C2 Forums. Of course not everyone will agree with each other on all points at all times. That’s what makes us human, and what makes life interesting, and what makes Internet forums like this one crackle with energy.
Matthew
Mathew,
If the performance graphs only show a specific dollar amount in history of $10,000 capital I am sure the vast majority of people would assume they could have traded the strategy back then with the same amount as is displayed, and with stock systems you could have. It doesn’t say “you could never have actually traded this unless you had $100,000.” If so, then subscribers would be warned that the volatility is actually much larger than a $10k account could handle and they would get wiped out if they tried.
But I see that you won’t budge on this, and it also goes against the profit motive for certain managers and C2’s bottom line, so we’ll agree to disagree and I’ll be more careful in avoiding any futures system that’s been re-scaled.
Thank you for considering the issue.
