Sharpe ratio calculation in MT5 is very unrealistic

To add comments, please log in or register
fxmz8008
21
fxmz8008  

I think the way MT5 calculates the Sharpe Ratio (SR) as explained here https://www.mql5.com/en/forum/1372/page2 is incorrect.

For several reasons:

a, It assumes a RiskFreeReturn of 0. Rate on 10 year US government bond is 2.23% (source: https://www.treasury.gov/resource-center/data-chart-center/interest-rates/Pages/TextView.aspx?data=yield)

As a consequence of this, you get unrealistic SR's (see next item). You might set it to zero for simplicity's sake, but that is not the way to go when presenting a well known ratio - stick to the formula. Besides that it also stays roughly the same for a year or so, so only a yearly update would already make it way more realistic.

b.  It doesn't take into account the amount of capital used - it assumes "all". That causes that it is irrelevant for the SR if you have achieved the results using 1% or 100% of the capital, which is of course nonsense. (If with only 1% you would also have fixed returns on the other 99% of the capital - see a.).

Besides the fact that the way of calculation is unrealistic, it leads (more importantly) to wrong conclusions: in the example two strategies can have exactly the same SR while the 1% capital used one is of course way better because, besides of returns, 99% of capital has been always safe. It should use the average used capital over the time period.

c. It doesn't take into account the time period.  Now two strategies can have the same result/risk, resulting in the same SR, while one generated that result in one month while the other one did it in a year. Would you consider these both as good?

Source: https://en.wikipedia.org/wiki/Sharpe_ratio

.


In short, the number you calculate this way is very far from the truth

Sharpe Ratio - is it correct?
Sharpe Ratio - is it correct?
  • 2010.07.26
  • www.mql5.com
MQL5 programming forum
Juer
182
Juer  
Any 3d party EA to calculate "realistic" Sharp ratio?
Sergey Golubev
Moderator
108436
Sergey Golubev  
As far as I remember - there are several ways to calculate Sharpe Ratio, and all those ways are correct ones.
fxmz8008
21
fxmz8008  

@ Sergey: I think the only correct Sharpe ratio definition comes from Sharpe himself, that's why I copied his revised formula.

If there were many correct ways to calculate a Sharpe (or any other) ratio, then it would be useless for any comparison.


@ Juer: I don't make 3rd part apps, but it's not so hard to calculate. What is more important is that you should be aware what you are comparing to what.

If you compare 2 strategies that have both used X capital and were over over time period Y, then YOU can use the "MQL5 Sharpe ratio" to compare two strategies.

But it doesn't say anything in comparing your "MQL5 Sharpe ratio" with others ratio, or having used other test variables like capital or time period tested over.

Sergey Golubev
Moderator
108436
Sergey Golubev  
fxmz8008:

@ Sergey: I think the only correct Sharpe ratio definition comes from Sharpe himself, that's why I copied his revised formula.

If there were many correct ways to calculate a Sharpe (or any other) ratio, then it would be useless for any comparison.


@ Juer: I don't make 3rd part apps, but it's not so hard to calculate. What is more important is that you should be aware what you are comparing to what.

If you compare 2 strategies that have both used X capital and were over over time period Y, then YOU can use the "MQL5 Sharpe ratio" to compare two strategies.

But it doesn't say anything in comparing your "MQL5 Sharpe ratio" with others ratio, or having used other test variables like capital or time period tested over.


I do not know about MQL5 Sharpe (about which one they took as a based one for the calculation), but I remember - I made one post about Sharpe many years ago:

Forum on trading, automated trading systems and testing trading strategies

Rent a Signal service development and suggestions

Sergey Golubev, 2009.01.12 07:47

Sharpe Ratio was added to advanced stats (thanks to Igorad).

Mathematics in Trading: How to Estimate Trade Results - MQL4 Articles

There are Sharpe Ratio, Modified Sharpe Ratio and Annual Sharpe Ratio. Sharp Ration telling us about stability (consistently) of the profit for some signals (lowest value compare with Modified Sharpe Ratio and Annual Sharpe Ratio) The Sharpe Ratio

Sharpe Ratio is not telling us about how profitable the system/signals. It is telling us about stability of the income.

For example,

- one signals did 0.1% for the 1st month, 0.1% for the second month and 0.1% for the 3rd month. Sharpe Ratio will be very hight. Because it is stable income every month. Same with banmks for example if you will deposit money to them.

- other signals did 1% return for the 1st month, -2% return for the second month and 5% for the 3rd month. Sharpe Ratio will be low. Because this signal provider is not giving stable consistent profit for every period of time.

For example, the signal ID 375 (by fxtrader625) is having Sharpe Ratio = 2.65 which is very good. If the signals are profitable so it indicates that it may be very good to subscribe.

So, the buyers should look at Sharpe Ratio (stable income or not stable income), than ROI, than max/min lot size, than how many pairs the signal provider is trading (it is difficult to receive big Sharpe Ratio for 1 pair during the long period of time trading) and so on.

Just for information.


So, according to my opinion - the Sharpe Ratio is very difficult to use in forex. For example, many people understand the following: "subscription to the signal = investing" ("I decided to invest my money to subscribe to the signal"). And they may ask: how much profit will I have monthly/yearly? 2% from deposit? 10%? or 20% yearly (20% every year) ... but we know that it is inpossible to estimate the fix profit from which an investment as the forum for example .... 

Sharpe Ratio may be good to use/to estimate in portfolio when we need the stability in income (small profit but same one every year for example).

Sergey Golubev
Moderator
108436
Sergey Golubev  

Yes, and this article from Sharpe: 

The Sharpe Ratio
William F. Sharpe
Stanford University
Reprinted fromThe Journal of Portfolio Management, Fall 1994
Sergey Golubev
Moderator
108436
Sergey Golubev  
If MetaQuotes will allow us to subscribe to more than 1 signal on same account (to create portfolio) so the Sharpe Ration will be the main indicator for us in this case.
Stephen Njuki
4961
Stephen Njuki  

fxmz8008:


b.  It doesn't take into account the amount of capital used - it assumes "all". That causes that it is irrelevant for the SR if you have achieved the results using 1% or 100% of the capital, which is of course nonsense. (If with only 1% you would also have fixed returns on the other 99% of the capital - see a.).

Besides the fact that the way of calculation is unrealistic, it leads (more importantly) to wrong conclusions: in the example two strategies can have exactly the same SR while the 1% capital used one is of course way better because, besides of returns, 99% of capital has been always safe. It should use the average used capital over the time period.


.


In short, the number you calculate this way is very far from the truth


Just because you are using only 1% of your margin to open a position, does not mean you are risking 1%... unless your leverage is 1:1
To add comments, please log in or register