How much would you risk on a 90% trading system

I feel the standard 2 % is way to low here,
that is, if the winrate would be 90%,
how much would you risk on each trade?

what is the risk reward?

Good question!

90% winrate doesn't mean nothing.

The system can be a loser with a win rate of 90% ...

The system can be a winner with a win rate of 10%

The important is the combination of the win rate and the risk reward ratio ...

The answer is simople and logical

I am going to assume you have 90% win ratio system, other wise you wouldn't ask it.


#1 It make logical sense that if you TP 3 pip and sl 100 pip. YOU WILL GET 90% win..

#2 Let's do the math if your starting balance is 1000$.. 90% win mean ..you win 9 trade and LOSE 1 trade on average.
So if you win 9 trade you win 3pip x 9 = 27 pip. Your 1 lose trade will be -100 pip.

So now if you do 50% risk that mean => you will lose 500$ in 1 trade.
And 500$ divide by 100 pip sl =equal=> 5$ per pip lot size.

So that mean if you bet 5$ per pip your 1 LOSS = -500$
BUT if you win then it's 5$ x 27 pip = 135$

Now that you get the basic of calculation.. NOW You can PLAN..
So if you win 9 trade and lose 1 trade you end up -500$ + 135$ = -365$ GUARANTEE... now no one is stupid to play this mm where if you win 9 trade and then 1 trade you lose more than your gain..

So once again you CUT down the risk and do the math AGAIN.

you choose 10% risk this time...
SO 10% risk of 1000$ = 100$ loss

To get lot size you do: 100$ divide by 100 pip sl =equal=> 1$ per pip Lot size

Now you you take: 1$ per pip x 27 pip = you win 27$

So summarize your result: you play 9 win trade you get 27$ .. one lose trade you lose -100$ => Once again you are GUARANTEE to lose -73$ every 10 trade..

So that is the LESSON HERE?
The lesson is your 90% win is CRAP IF YOUR SL IS TOO BIG... NO matter how much SAFE RISK you do even at 1% or -0.01% YOU WILL ALWAY LOSE...

-------------
Lot of of newbie who focus on small risk % alone never know why they never win even if they play safe..risk. Now they know..

You need to balance out Reward TP value to close to SL AND STILL ACHIEVE 90% win rate ... if you can do that you own the holygrail....

Good luck..

I would go with at least 1 : 1.5 / 2 R:R. This way you can profit more from your 90% system.

Maybe your system cannot accept 1:2 ratio. This would be a problem, and as FerruFx said, it can be a loser even if it's a 90% system.

As a general rule, increasing the RR will decrease the win rate in similar proportion.

@roughtrader: The table here shows the probability of losing X consecutive trades in any 50 trade sequence, for any given win rate.
However, unfortunately, it does not show the more important probability of the losses exceeding the wins by a given number. For example, if you have a sequence of LLLWLLL, that is a max sequence of 3 losses, but in reality, you've lost 6 out of 7 trades (hence if you were risking 5% per trade, your account would be down (6-1) x 5 = 25% for fixed percent, or 100–(100 x 1.05^1 x 0.95^6) = ~ 22.8% drawdown for fixed frac — all assuming 1:1 RR).

The risk taken on each trade should take into account all the variables at play in your trading system.

That is to say; your capital, the win rate of your system and the average risk reward ratio that your trading system gives to you.

It's amazing how many times you read here and there that you shouldn't risk more than 2% of your account on any single trade and that is just plain wrong as this of course would be too low for a trading system with a good win rate and providing trades with a good risk/reward, but also too much high on the contrary case !


Here is a formula for you to play with it that calculates the optimum risk size for any trade according to your win rate and risk reward ratio.

B1 : Percentage of wins [%].
B2 : Percentage of losses [%].
B4 : Amount of wins [%].
B5 : Amount of losses [%].
B6 : Risk Factor (0,3).
R1 : Possition Size in percentage of your capital [%]

B1 and B2 are from your win ratio, in your case B1=90, B2=10 as you said you had a 90% win ratio trading system.
B4 and B5 are the average the risk reward of your last trades. This means the average you won and lost as a percentage of your capital.
B6 is an arbitrary value : 0.1=very conservative, 0.2 = conservative, 0.3 = optimal, 0,4=aggressive, etc, a word on this later...

Finally, R1 is the calculated position size as a percentage of your account capital.

For this to be accurate, you MUST have a valid trading history giving an accurate win ratio and average RR of at least 100+ trades, is that's not the case, then adjust the Risk factor to 0.2 or even 0.1 until you have accurate data from real trades.

Here is the formula:

R1 = ( (B1*B4)-(B2*B5) ) / B4 * B6

Example for a system which has a win ratio of 60% and an average Risk/Reward of 1:2
B1 = 60
B2 = 40
B4 = 2
B5 = 1
B6 = 0,3

R1 = ( (60*2)-(40*1) ) / 2 * 0.3
R1 = ( 120-40 ) / 2 * 0.3
R1 = 80 / 2 * 0.3
R1 = 40 * 0.3
R1 = 12

So if you have 1000$ in your account, you can risk 12% of it on next trade, this is way too much isn't it ? Then adjust risk factor...

Play with it on an excel sheet and see the results...

BR/

Cool formulas

One just got to remember that the very first trade can be the one which falls into 10%.

wow 82% chance of 11 losses in a row for 25%? Is this accurate? Thanks for the table, I saved it.

It is just a rewriting of the well known simplified Kelly criterion
http://en.wikipedia.org/wiki/Kelly_criterion

While we're at the Kelly, do you have an online calculator or an excel spreadsheet I can use? The site I used went offline and I'm clueless with excel

Certainly, it's just that I have them since long time and don't remember from which book I took them (Read to many books that are now filed away )
Thanks for remembering

Go download "Open Office" if you want to have excel ....

Agreed.

What you've described looks very much like the Kelly formula (here, here), multiplied by a risk factor (B6).

The Kelly is considered to be too aggressive, because while it delivers the optimal bet size for maximizing returns, it doesn't take risk of ruin into account (assumption: a trader's overriding priority should be capital preservation). I assume that the risk factor (B6) is an attempt to solve this problem. However, multiplying the Kelly value by an arbitrary value merely creates another arbitrary value.

For instance, in your example the Kelly value is 40%. Now you can multiply this by 0.01, 0.05, 0.1, 0.2, 0.25, 0.3, 0.4, or anything else — i.e. assign any value you wish to B6 — to obtain almost any R1 value that 'seems' about right.

[EDIT] Looks like PipMeUp's and my posts crossed.

It's simply the result of a binomial probability calculation (see formulae in attached XLS).

I would say that this question is very hypothetical. Reason being is because you have many different ways to look at it. A pessismic will say "If you win 9 out of 10 times, your 1 loser could wipe our your entire account". That of course is true, but also depends heavily on the trader himself. It would also depend of the trader is compounding or not. Imagine if you win 90% of your trades, and on the 10th trade you are compounding, and that is the losing trade.

I would say the win rate isn't important at all. What would be important (to me) is your ACTUAL r:r. The expected R:R means nothing, because most trades move their sl when it is about to get hit, or when they stop out, they take the very same position which caused them to stop out in the first place.

The amount of pips on avg (Pip-drawdown) you give to the market is by far the most important stat to me (as a scalper). The less pips you give to the market, the more the system is worth to me. If I can give 5 pips to the market before I lose 5 pips then that would be worth GOLD to me, then winning 90% of my trades.

ill put it simple from a personal point of view, if the trade that im about to take is a 90% winner, then I dont mind risking 50% or more. im just talking about a sible trade, not so sure about systems that has that winning percentage,

I like this question...

So we are now talking about a fixed R:R...

Let's go into the numbers...

Let's take an example in which you have the reward twice as big as the risk taken...

First trade : TP : 10 pips / SL : 5 pips
Second trade : TP : 100 pips / SL : 50 pips

No need to make complex calculations to find out that even though you have a nice system (based on stats) on the paper, you might end up losing a lot in this... Both trades have TP twice as big as SL (which respect the given RR), but if you win the first and get SL hit on the second, you will be a big time looser...

In your system, in addition to the RR and Winrate, you should also take into account the distribution of the SL and TP (with historical data for backtest)...
It seems like a lot of calculations for a "sure" system...

That assumes that the position size, in absolute terms, on each trade is the same, e.g. 1 lot.
But if I was to risk (say) 2% of my account (fixed frac sizing) on each trade — as many textbooks recommend — then the risk in both trades is equal.