Leverage, Test results, & risk

Posted By: DdlV

Leverage, Test results, & risk - 07/28/14 15:44

Hi. Taking a look at leverage, Test results, and risk. Using Z5 since it only trades 1 asset.

Using this 1:200 AssetsFix.dta (EUR/CHF Margin=8):

Code:

EUR/CHF   1.21554 0.00025 0.0400 -0.1300 0.0001 0.08228 8.000 1000.0

Z5 Tests to:

Code:

Z5 (oP group) ..
Z system 5.25
Ini: MMax 100  Wkend 2  Verb 2  

BackTest: Z5 EUR/CHF 2011..2014
Worst Case Capital   2914$
Monte Carlo Analysis... Median AR 157%
Profit 10095$  MI 295$  DD 941$  Capital 1781$
Trades 2482  Win 95%  Avg +10.9p  Bars 112
AR 199%  PF 9.97  SR 2.02  UI 4%  R2 0.63
Time 00:02:23

Using this 1:400 AssetsFix.dta (only change is EUR/CHF Margin=4):

Code:

EUR/CHF   1.21554 0.00025 0.0400 -0.1300 0.0001 0.08228 4.000 1000.0

Z5 Tests to:

Code:

Z5 (oP group) ..
Z system 5.25
Ini: MMax 100  Wkend 2  Verb 2  

BackTest: Z5 EUR/CHF 2011..2014
Worst Case Capital   4983$
Monte Carlo Analysis... Median AR 193%
Profit 20159$  MI 589$  DD 1959$  Capital 2824$
Trades 2482  Win 95%  Avg +10.9p  Bars 112
AR 250%  PF 10.11  SR 2.01  UI 4%  R2 0.64
Time 00:02:26

Worst Case Capital, Drawdown, & CR are of course much higher given the doubled leverage.

For 1:400 leverage, the Margin setting that gives closest to the original CR above is 32:

Code:

Z5 (oP group) ..
Z system 5.25
Ini: MMax 100  Wkend 2  Verb 2  

BackTest: Z5 EUR/CHF 2011..2014
Worst Case Capital   3185$
Monte Carlo Analysis... Median AR 193%
Profit 13010$  MI 380$  DD 1250$  Capital 1801$
Trades 2462  Win 95%  Avg +11.1p  Bars 109
AR 253%  PF 10.88  SR 2.04  UI 4%  R2 0.66
Time 00:02:15

Comparing the above to the original 1:200 leverage, there are slightly better Worst Case Capital, PF, SR, and R2 (probably not statistically significant?); and rather better AR (~+25% - both the single Tested figure and the MC Median). So, higher leverage appears better.

However: The manual says "When trading the same number of contracts with the same budget, a low leverage account is always more likely to crash than a high leverage account." which makes sense. However, with the higher leverage account Zorro trades more lots per trade, so even though the CR (budget) can be set equivalent via the Margin slider, the statement above doesn't apply since the contracts have increased. It seems instead that for the same CR while the risk of Margin Call might be the same in terms of the single-number probability of "actually crossing the limit line", with the higher leverage account the speed of approaching that limit would be greater and hence the time to react less. Also, it seems this probability could actually be greater with the higher leverage since the effect of volatility would be magnified? Is this the correct way to view things?

Thanks.

Posted By: jcl

Re: Leverage, Test results, & risk - 07/29/14 08:36

Higher leverage increases the AR because the required capital is the sum of drawdown and margin.

It is hard to tell whether the margin call risk is increased or decreased by using the capital that was set free by the reduced margin for buying more contracts. The equity goes down faster, but the balance grows faster. I would say that the short-term risk is higher but the long-term risk is lower.

The difference is normally not very significant because in most strategies the drawdown is much higher than the margin.

Posted By: DdlV

Re: Leverage, Test results, & risk - 07/29/14 11:51

Thanks jcl. Except...

CR has been adjusted to the same via the Margin slider, and yet the AR is greater, presumably because for the same CR the strategy is now able via the lower margin needed to do larger lots of its (on balance) profitable thing?

Thanks.