Surely a high Scaling Factor % (in comparison to trading account size) is unwise (especially for a newbie like me).
_____________________________________
But help me out here. My calculations show:
Scenario A - Conventional:
If I autotrade a $30K account using a Forex trading system with a 30% Scaling Factor, and the system returned 46% in one year, I could well make around only 23%* real profit (before tax).
[assuming: $1/trade BulldogFX cost; 1.2trades/day; 7day av.trade length; 20% trade slippage; 1%/year average interest rate on rolling currency; 300days active trading; $250/m subscription fees]
In doing so I am risking $30K if the system blows up, although in this case the trading system would need a C2-reported drawdown of 85%* before ManFX liquidated my account.
[assuming: Max 45% account capital in play before the drawdown; 100:1 ManFX account leverage vs. 33:1 C2 account leverage]
---------------
Scenario B:
What if I used the same Scaling Factor, but risked less money (as suggested by Science Trader I think*)?
[*** http://www.collective2.com/cgi-perl/board.mpl?want=listmsgs&boardid=9323885&threadhilite=3820&displayblock=36_7&sorttype=alpha&session=344559354864231412444701705333915]
This time I’m autotrading a $15K account with the same 30% Scaling Factor, and if the same system made 46% I could make (by coincidence) 46% (before tax).
Plus, I’m only risking $15K this time.
The downside? Well, this time, the system only needs to drawdown 35% (as reported on C2) before my account expires.
________________
So what’s it to be?
With Scenario A, I can weather a storm more easily, but I’m risking twice as much capital and I have to accept far lower profits unless I go for a riskier trading system.
With Scenario B, I can’t take on the riskier systems, but I can get better returns anyway (albeit on the smaller capital), so can look for a safer system (as I think Jules suggested in the advanced stastics introduction where he explains how to leverage the Sharpe value). Plus, if it does blow up I only lose half the money compared to that I did in Scenario A.
So, which strategy is overall less risky?
Set me right!
Charles
you don’t need to agonize over this one. If you are a newer trader, do not use significant leverage for quite some time. The prevailing belief is that 90-95% of new traders using leverage (forex, futures, options mostly) stop trading in the first year or so (usually due to losing their trading capital).
Secondly, the perceived possible drawdown of a system has almost no meaning. “Black swan” or “unexpected market events” or “oops, didn’t see that one coming” is what you REALLY have to prepare for. People with money in their mutual funds/401Ks generally recover fine over time. Leveraged traders, it is called “Risk of Ruin.” Systems never bother with this. They trade away until they suddenly stop publishing due to a wipeout. Then they pick themselves months later and start a new system. Notice how some vendors have other systems. Some of them look like a “black” swan dive…
New traders are like new drivers - they have many more accidents on the way to trading maturity. They have a very very high death rate.
So your question:
Do you want to lose all of your money within about 18 months, or do you want to pursue a 23% return?
Hi Ross.
Thanks for the response. I completely agree with your sentiment. In the way you have stated your argument, the conclusion is obvious: Unless I can convince myself otherwise, I will default to a low-leverage system.
… On the other hand, my question was a bit different:
Do I want to risk half my money in about 18 months? Or risk losing everything for just 26% return?
Maybe you are right… maybe I should beef up my mutual funds instead. Some of them (Latin America) are exceeding 26%… This was all going to be so fun :0(
just trying to spare you the route taken by "the 90%" Everyone thinks "gee, at 600% per annum, I could have 10 million in 18 months" as advertised by the Ebay sellers…"
it takes time to learn. There are no shortcuts, but we tried to give you one here. Preserve your trading capital, dont flush it down the toilet. Spend a couple years learning, so you can become a serious trader. Dont start by throwing away your trading funds first
Good thinking. The problem is that judging the exact risk of a system is not possible. The Sharpe ratio, historical drawdowns etc. might give a first indication, but generally the number of observations on which they are based are too short to be reliable exact forecasts of future performance. Furthermore, these ratios are often not stable over time and might change as a result of a sudden increase in market volatility. A good example of the former is TMG. It looked like an extremely low-risk system until the recent drawdown, and as you can see from some early reviews, subscribers traded it with more leverage (using options or leveraged Proshares QLD instead of QQQQ) than shown on C2. Rocket Science-Mini Dow appears to be a good example of the latter, where an unexpected increase in market volatility caused some sudden wild swings.
Another thing you have to be aware of is that you can’t always simply multiply the returns by a certain leverage level to obtain the leveraged returns. E.g. an unleveraged system with 50% annualized return, might give you 150% (instead of 100%) annualized return when using 2x leverage. Similarly, a system with 100% annualized return that used 2x leverage might only give you 30% (instead of 50%) annualized return without leverage. This effect is due to compounding.
In scenario B, you say that the profit is 46%, but this is (if I understand you correctly) 46% of the 15K account, which is still 23% of your total capital. So the profit in $ would be approximately the same; only the risk has a different profile.
I’m not so sure that you cannot loose more than $15K in scenario B. I’m not sure, but I think this can happen too: Suppose you buy $200K of currency X. This means you have a debt of $185 to your broker. Suppose the value of currency X decreases by 50% overnight because some European banker sneezed, then after opening you posses only $100K worth of currency X, but you still have that debt of $185K. The broker will close your account, and then you will still owe him $85K. This scenario is not very probable, but is it impossible?
Anyway, let’s assume that you’re right and that you cannot end with a negative account. Then the maximum loss of scenario B is smaller, but the probability that it will happen is larger. So you cannot say which scenario has the most risk unless you give an exact formula for what you mean by “risk”. Furthermore the probabilies of the various events have to be known before you can calculate risk.
Scenario B is comparable to using a tighter stop loss rule. While it is true that a stop loss can protect you from large losses, it is also true that a too tight stop loss rule increases the probability of a loss. For example, you could do it more extremely and have only $1000 on your account while using excessive leverage. Then you will loose no more than $1000, but there is a very large probability that you will loose it, so it can still be foolish to do.
Note that in either case you are not obliged to trade the system untill your account is empty. You can stop much earlier, e.g. after a 30% drawdown on C2. In that case I think both scenario’s are equivalent except for interest.
Hi ScienceTrader,
"The problem is that judging the exact risk of a system is not possible.“
This is an excellent point. Unless I can quantify the risk of two systems with some confidence, and factor that into a probable drawdown value (perhaps one of Jules’ Expected Shortfall statistics), then it’s all for the cat. I guess you’re right - all those statistics are based on historical data; usually over a time-period which is small in relation to the time-period of fundamental changes in the Market dynamic.
Jules, how should we use your advanced statistics then? Perhaps as a better / worse indicator between systems, but not as a risk quantifier?
”…you can’t always simply multiply the returns by a certain leverage level to obtain the leveraged returns."
Of course you’re right. Be assured that the numbers from my spreadsheet came out as nice and round by coincidence. Of course, they were all entirely hypothetical; if your point is that you can’t base a conclusion on a narrow choice of variables, then quite right, I absolutely accept that.
Thanks,
Charles.
Hmmm… some good answers in here.
"In scenario B, you say that the profit is 46%, but this is (if I understand you correctly) 46% of the 15K account, which is still 23% of your total capital. So the profit in $ would be approximately the same; only the risk has a different profile. "
Correct. But, of course it frees up the other $15K for me to go and make use of.
"I’m not so sure that you cannot lose more than $15K in scenario B.“
An obvious point! How quickly does a Forex broker such as ManFX act to liquidate open positions? That’s crucial to know. I’ve never been caught out on my demo trading in the UK to date. Anyone had any experience of problems?
”…you cannot say which scenario has the most risk unless you give an exact formula for what you mean by ‘risk’.“
True. In this case I’m thinking: (e.g.) ‘% Likelihood of Running Out of Margin’. In the case of scenario A, this could be (in words) “The Chance of an 85% Drawdown using System1” (to statistical confidence of e.g. 1.96 Sigma). Similarly, in Scenario B it would be “The Chance of a 35% Drawdown using System2”. Are there any statistics that could help me there?
”…a too tight stop loss rule increases the probability of a loss."
Yes, understood (when compared to a looser stop loss rule being traded with the same system).
"You can stop much earlier, e.g. after a 30% drawdown."
Good point. Of course I’d still be living with lower profits, but at least I wouldn’t be risking all of $30K. Perhaps that’s the answer.
-----
Hmmm… in that case, realistically I could still probably get better returns elsewhere…
I must get home. Dinner is in the dog, no doubt.
Thanks all,
Charles.
There are 2 additional concerns with high leverage:
1. The possibility of a failure of technology. Systems with high leverage and low risk typically hold positions for a short time (minutes, hours). If you or your auto trading software somehow is not able to close a position once a signal is issued, your account can be wiped out fast.
2. Widening spreads in fast moving (forex) markets. An (un)expected news event might lead certain forex brokers to suddenly increase their spreads for a short while. If you’re leveraged heavily this can cause a wipe out as well.
Charles,
The advanced statistics are also intended as quantifaction of risk. But it can be quantified in different ways (e.g. VaR, Expected Shortfall, probability of ruin), and there is no single standard to which everyone agrees. Basically this is because there are always at least two different aspects of risk, namely “probability of a loss” and “size of a loss”. There are many ways to integrate them into a single risk measure. This is why I said that you can determine the risk only if you choose a formula first.
But suppose you choose to base your decision on Expected Shortfall. Even then these statistics pertain only to the system as it is, and not some alteration of the system that you make if you implicitely add some stop loss rule by having a higher leverage with a smaller account. Perhaps you can compute the expected shortfall for that case yourself, but then you will at least need the various probabilities that are involved. So without further specification of the properties of the system, the question can’t be answered.
So if you see that the system has a very low risk (according to the definition you chosed) without leverage, you can use that as a reason to increase leverage, but note that you won’t know exactly what the risk of the leveraged system will be, unless you make further computations.
The theory of Sharpe that you can increase leverage for systems with low risk assumes that the broker will not interfere. So either you borrow the money elsewhere or your leverage should be such that your account stays far away from liquidation.
Apart from that it is true what Science Trader said, that the estimates are often unreliable. Personally I think that one of the main advantages of the advanced statistics is the confidence interval for the Sharpe ratio. In many cases it makes clear that after a few months trading the estimates are still very unreliable and that you basically know nothing about the profitability of the system.
A related problem is that statistics ideally are based on a representative sample from a larger universe (in this case of returns). A confidence interval (obtained from the sample) should give an idea about the possible values of the statistic in the entire universe. But in the case of C2 the sample is most likely not fully representative of the larger universe.
For example, over the past year or so, we haven’t seen many extreme market conditions. Over the last year monthly S&P 500 returns were in the range of about -3% to +4.5%. But historically more than 30% of the time S&P500 returns fall outside this range, with extremes that are an order of magnitude larger (-21% and +16%). And it’s an open question how much information the current statistics give for a system’s risk and reward in these extreme cases. The same might apply to forex.
Jules said: "I’m not so sure that you cannot lose more than $15K in scenario B. I’m not sure, but I think this can happen too: Suppose you buy $200K of currency X. This means you have a debt of $185 to your broker. Suppose the value of currency X decreases by 50% overnight because some European banker sneezed, then after opening you posses only $100K worth of currency X, but you still have that debt of $185K. The broker will close your account, and then you will still owe him $85K. This scenario is not very probable, but is it impossible?"
Out of interest, and just to add some colour to this: I am thinking of autotrading (server-based) through BulldogFX, so I asked the broker, ManFX in this case about the possibility of negative account value, and this was their response to me:
"Let us say you have $25,000 in an account, and have a position long 200K of EUR/USD at 1.3500. Assuming a margin setting of 1%, you are putting down $2000 for this trade as used margin, and have $23,000 remaining as useable margin. When the floating losses on this trade exceed $23,000, your useable margin will be below zero, and the margin call will occur immediately. In this example, that would occur when the EUR/USD fell below 1.2350 (200K of EUR/USD is $20 per pip, 23,000/20 = 1150 pips). If the margin call was filled at the price it was triggered at, you would be left with just under $2000 in the account (which was your used margin).
"Let us say that instead of a 200K position, you opened a 2 Mil position long EUR/USD at 1.3500. That would mean $20,000 as used margin (on 1% margin), and $5,000 as useable margin. At $200 a pip, it only takes a 25 pip move against you to trigger a margin call (5000/200=25). If the margin call was filled at the price it was triggered at, you would be left with just under $20,000 in the account (which was your used margin).
"Negative balances can occur, and there is no maximum negative balance. That having been said, they are exceedingly rare and usually small relative to the size of the account. In both examples above, there is a risk of a negative balance. As you can see, negative balances can occur if the margin call is filled at a less favorable price than where it was triggered, this is more likely to occur in volatile market conditions."
Which seems quite sensible, and is slightly reassuring to me. I interpret this as that the risk is still there, but managed to a high degree.
Charles.
Thanks for checking. That sounds like what I suggested: It is improbable, but not impossible. The example of EUR/USD is somewhat biased, because both currencies are based on large economies. What if the system suddenly starts to trade Thai bath (or whatever they have there)? Perhaps the margin requirements are different for that too, but then you leverage schedule won’t work anymore.
Anyway, I don’t want to argue that this is too dangerous, but the facts should be clear. The more important point is that I think that your decision to stop trading a system should be based on an assessment of the future profitability of the system, and not on some arbitrary leverage & margin construction. So if you loose your faith in the system after a loss of 30% then you should not continue trading because there is still some money left in the account, and conversely your leverage should be set such that you are not forced to stop trading after a loss of only 10%. That is, first you decide about your “stop loss” rule, and only then you determine the leverage.
