Menu
Sister Sites

Options

Bookmark Thread

First Page First Unread Last Page Last Post

Print Thread

Similar Threads

What is your win rate and risk reward ratio? 62 replies

The Smoke and Mirrors of the Risk/Reward Ratio 22 replies

Rationale behind "Risk Per Trade" and "Risk/Reward Ratio"? 8 replies

Risk/Reward Ratio and the power behind account growth 15 replies

Good reward/risk ratio experiment 20 replies

Attachments: Risk Reward Ratio and Account Risk
Exit Attachments

Risk Reward Ratio and Account Risk

Thank you for your answers!

Quoting radicalmoses
Hello Dottore,

2) The model concerns his AR with respect to RRR. No matter what % of account he is trading,, his next trade probability is always 50%. And if it turns right, its due to the system's probability (win/loss ratio)which can be 60or70% or whatever.
The concept of probability comes from empirical results. If a coin is symmetrical you have 50:50 chance (large sample of tosses) but if it is not you can have 70(head):30 (tail) (large sample of tosses) for example.

So the question is: if you toss flawed one what will be the chance=probability that there will be a tail?

And the second one: is fx comparable with first or second?

and why?

Quoting radicalmoses

3) The model does not concern the drawdown on the system. No matter what the system, you will at some point suffer drawdown in your account and hit a losing streak. The losing streak of any system is a variable. It can be say 4 trades in a row. so suppose the model tells you to trade 4% of the account on the trade and you hit a losing streak 4 times in a row, thats 16% down already. Very possible scenario.
The strength of any MM model is not only in containing risk but also to cause geometric growth which is only possible in leveraged instruments.
If only risk was the criteria, might as well put money in the bank and earn interest.
Just some of my thoughts to put things in perspective from what little I know about MM. Its a vast subject and open to debate for ever.
Drawdown on this stage of analysis is not taken into account.

The strangth of this model lies in fact that it adjustes volume of entry to expected )probable) market conditions, and drawdown is a step further in mm. In fact (if we can prove it by backtest) that his method actually works good it gives an additional edge: decreases risk and increases gains (yes - geometrically by compounding or linerarry or whatever). But ASai asked in his first post what we think about it as a tool to mitigate risk.

have a nice weekend!!!!
Quidquid latine dictum sit altum viditur
I quite agree with you. I actually decided to stick with fixed VAR for the moment. I'll definitely backtest this system in the future and if it works, I'll switch to this system.



Quoting dottore
I finally understood your post , forgive me straightforwardness as I’m still having problems with my English.

Here is a main problem I spotted:

“This method reduces the lot size when the loss size increases and increases the lot size when the loss size decreases so it never causes larger than acceptable drawdowns.”

In fact “loss size” is settled when you enter the trade and comprises two parts – SL and VAR.
And you cant simply reduce the loss AFTER entry (of course you can take a portion of a trade
if market goes against you, I can think about usability of this approach)

So your goal is to decrease the risk.

Understanding it as a value at risk I do not know better method than decreasing VAR.

On the other hand – I’m not sure if common def of risk always works. Consider situation when you have high risk/high
reward system and you gradually take a portion of account – and in the end (month, quarter) you have a better outcome
(far better) than risking 2% in regular account. And this is not an academic discussion, I saw and analyzed such systems and their usefulness.

Back to your system:

As I understand you are estimating probability of winning/loosing by calculating relation of “probable gain” to account at risk (VAR)
And you are adjusting size of your position.

1. I believe that is a good method but it would be nice to test it. Could you backtest it for example on 200-300 trades?
2. The one possible problem is with the method you use in calculating T/P – you assume it is reliable.
3. Do you know any other methods of calculating probability of winning in next trade?

We should work on a possible simple test to verify reliability of Risk Reward Ratio as a good estimation of probability of winning in the next trade:
The best one I believe will be a backtest, I will try to figure how to do it a little later, as I have work to do now.

Take care
Quoting aparsai
I quite agree with you. I actually decided to stick with fixed VAR for the moment. I'll definitely backtest this system in the future and if it works, I'll switch to this system.
Thanks for your answer - I'm actually interested in systems with flexible VAR and as I understand the key here is a good method of "predicting" the next day's move. Please if you can - describe briefly your metod of calculating TP. I can backtest it myself. Or you can PM me.

Thanks!
Quidquid latine dictum sit altum viditur
Quoting dottore
Thanks for your answer - I'm actually interested in systems with flexible VAR and as I understand the key here is a good method of "predicting" the next day's move. Please if you can - describe briefly your metod of calculating TP. I can backtest it myself. Or you can PM me.

Thanks!
In most cases I use Fibonacci retracement to calculate the profit target.
hi guys,
i think that the best way to calculate profits is using the RRR and AR as described in this thread.

i also made some calculations about compounding the gains and i found that starting with 10'000$ with a 2% Risk and a 50% RRR you need 126 winning trades to achieve 1 Million $$$. (probaly 18 to 24 months)

and with a 33% RRR it will take about 82 trades. about a year.

so what are your comment on this theory (as more experienced traders)
Quoting nicolas_nrjk
hi guys,
i think that the best way to calculate profits is using the RRR and AR as described in this thread.

i also made some calculations about compounding the gains and i found that starting with 10'000$ with a 2% Risk and a 50% RRR you need 126 winning trades to achieve 1 Million $$$. (probaly 18 to 24 months)

and with a 33% RRR it will take about 82 trades. about a year.

so what are your comment on this theory (as more experienced traders)
Could you explain how you made those calculations?

Thanks in advance
  • Post #27
  • Quote
  • Edited 9:15am Nov 5, 2006 9:10am | Edited 9:15am
in a simplified way if your risk is 2% and RRR 50% your winning are of 4% per trade so:
=10000*1.04^118 (that is 118 times *1.04)
copy paste this equation to excel and you'll get the result

for the 33% RRR means 6% winning per trade
=10000*1.06^80

find attached a file to calculate how many lots to buy on each account size
Attached File(s)
File Type: zip RRR.zip   6 KB | 438 downloads
Your calculations are correct but you didn't take something very important into account, the "winrate". Most systems that have a RRR < 50% also have a winrate of less than 50%. That means they lose over 50% of the time. If you could have 118 winners without a loss then you could make such amount.



Quoting nicolas_nrjk
in a simplified way if your risk is 2% and RRR 50% your winning are of 4% per trade so:
=10000*1.04^118 (that is 118 times *1.04)
copy paste this equation to excel and you'll get the result

for the 33% RRR means 6% winning per trade
=10000*1.06^80

find attached a file to calculate how many lots to buy on each account size
you are right i did say that you need 118 winning trades in a row

but concider this:
21 trades per month with 14 wins 7 losses and a RRR=50%
that would be 14 - (7/2) = 10 winners per month
in 12 months you got your 118 winners
Quoting nicolas_nrjk
you are right i did say that you need 118 winning trades in a row

but concider this:
21 trades per month with 14 wins 7 losses and a RRR=50%
that would be 14 - (7/2) = 10 winners per month
in 12 months you got your 118 winners
Do you have such a system in hand?
  • Post #31
  • Quote
  • Last Post: Feb 22, 2007 11:37am Feb 22, 2007 11:37am
Suppose you win 100*x% of your trades, lose 100*y% of your trades, and break-even on 100*z% of your trades...and let's say that when you win, your account increases by A%, if you lose your account decreases by B%, and of course if you break even, your account does not change.

The expected return, is what return you can expect on average. This is calculated as:

E(return) = x*A - y*B

For example, if you win 45% of the time and when you win, your account grows by 8%, and if you lose 50% if the time (so you break-even 100% - 50% - 45% = 5% of the time)...and when you lose you lose 4% of your account, then your expected return is:
E(return) = .45*8 - .5*4 = 3.6 - 2 = 1.6%
Because this is positive, you can expect your account to grow by 1.6% on average. Assuming you always adjust your position size so that you are risking the same percent of your account on each trade (I call that geometric money management), then you can project the future value of your account: e.g. in 30 trades, you account should be up by, (1.016)^30 = 1.61, you should be up 61%.

What if you have $1000 and you want to be a millionaire in one year? Well then you need to grow your account 1000-fold. What is the expected return needed to acheive this growth if you make a trade every day. Let's say there are (52 weeks)(5 days/week) = 260 trading days in a year. Then we solve: x^260 = 1000 ==> x = 1000^(1/260) = 1.027. So you just need to average 2.7% return on your account per trade.

Playing with this "how rich am I going to be" math is fun! Now let's develop a strategy that actually achieves this mean return.
0 traders viewing now