Risk Reward Myth

[tommog]

Hi all,

I just thought I'd raise this (maybe controversial) point. All new traders are taught have your winners bigger than your losers. Thats fine, im not saying this is wrong, but I feel the logic is flawed.

The argument supporting this is you are rewarded twice as much for every time you are right (in a 1:2 risk reward example). However this implies your odds of having a winning trade are 50/50. They are not.

Assuming no slippage and commission (an ideal world) and you have a 1:1 risk reward profile the odds of you winning are 50%. You can flip a coin randomly in the market and you would win 50% of the time. After a couple of hundred trades flipping a coin you would be break even on ticks (again assuming no trading costs).

HOWEVER, if you risk 1 to make 2 the odds are not 50/50 because you are expecting the market to move twice as far in your favour to make a profit. To use the coin example again, 1:1 risk reward is a 0.5% chance of winning. However a 1:2 risk reward the odds of you winning are 0.5 x 0.5 = 0.25% chance of winning.

So again if we were to trade randomly following a strict risk reward of 1:2 (with no slippage and commissions) and you entered long on heads with a target of 12 ticks and a stop of 6 ticks (and vice versa) after a couple of hundred "spins" you would have won about 25% of the time and be about flat on ticks.

So, lessons I feel can be learned, if you are a new trader provided you stick with exactly the same risk reward profile on every trade (so the probabilities balance out) you do not have to have winners bigger than your losers. You will not make anymore money but I feel it is easier to stick to a plan that results in a winner every other trade than one that results in a winner every third trade. It will help confidence.

Also, this does not teach you how to have an "edge" in the market, thats up to you, but with a consistent risk:reward profile and consistent position sizing even random trading can keep you from quick blow outs.

Of course we have trading costs, so a 1:1 risk reward in theory is 50%, but after costs its more like 45% expectancy. Hence this is a negative sum game. However, if after years of studying the markets you can improve on randomness by 6% (not sure how this is expressed in mathematical terms) you now have a positive expectancy and are a profitable trader!

Please feel free to correct me if I am wrong, just highlighting its about expectancy not risk:reward which is where most people start (and stop) their learning. The greater your risk: reward ratio the lower your win rates will be. Why harp on about the risk 1 to make 3 all the time?

 

[shadowninja]

Well, it's like the time = money thread. It depends on the individual. They need to have some way of profiting from the market and it may be via a high probability, low reward vs risk or a low probability, high reward vs risk strategy but as long as they're profitable in the long run, it doesn't matter.

 

[Scotty2Cues]

The argument supporting this is you are rewarded twice as much for every time you are right (in a 1:2 risk reward example). However this implies your odds of having a winning trade are 50/50. They are not.

i dont think this is correct unless ive misunderstood. If your risk:reward is 1:2 then it means your trades need a success rate greater than 33% to profit and your r:r dont determine you winning odds

 

[HowardCohodas]

Mathematical Expectation
Gambler's Ruin

 

[Shakone]

I think for psychology reasons you want to have a higher than 50% win rate. If you don't combine that with approximately a 1:1 risk reward or better, you're likely to struggle imho. Sure you can have a 90% win rate, and your losers much bigger than winners. But a 90% win rate is hard to keep up. And if you're off your game, or the markets start behaving unusually...you'll lose a lot. For me at least, risk reward is hugely important.

 

[Mr. Charts]

Indeed, a beginner's mistake is to assume there is a 50/50 chance of success and is predicated on the trade situation being simply random.
They also assume that there is a 50/50 chance of a 1:2 or 1:3 etc risk/reward being hit.
That is absurd, but even seasoned traders have it rooted in their minds.
Richard

 

[HowardCohodas]

I think for psychology reasons you want to have a higher than 50% win rate. If you don't combine that with approximately a 1:1 risk reward or better, you're likely to struggle imho. Sure you can have a 90% win rate, and your losers much bigger than winners. But a 90% win rate is hard to keep up. And if you're off your game, or the markets start behaving unusually...you'll lose a lot. For me at least, risk reward is hugely important.

You are hugely correct about win rate being important to the trader's psychology. Some of us are constructed to require a high win rate in spite of what the mathematical expectation calculations show. I'm afraid I resemble that description.

There are quality assurance techniques to detect failure of the trader and/or failure of the strategy early enough to prevent substantial losses. The key element for implementing the technique for me is a daily honest journal.

 

[meanreversion]

I actually think risk:reward is a load of b0ll0cks myself, as it implies that the market actually cares about what you're doing. Furthermore, it implies that the only strategies which can make money are those with a ratio of 2:1 or 3:1.

You set the risk, the market gives you the reward.

However I think there's a little grain of truth behind the idea of 2 or 3:1, as it clearly implies small losing trades and big winning trades.

One of the main cognitive biases leading to losing money is cutting winners too soon and running losses for far too long. If you were to stick with 2:1, that's not going to happen.

Having said all that, I trade a trend model where I don't have any profit targets.. I've been unable to find a robust mechanical system which uses profit targets, and indeed when I did trade this way, I frequently came across the dilemma whereby (e.g.) the profit target was 100 pips, and the market would get to +97 pips before reversing. Profit targeting doesn't really tell you what to do in these circumstances.

 

[meanreversion]

I think for psychology reasons you want to have a higher than 50% win rate. If you don't combine that with approximately a 1:1 risk reward or better, you're likely to struggle imho. Sure you can have a 90% win rate, and your losers much bigger than winners. But a 90% win rate is hard to keep up. And if you're off your game, or the markets start behaving unusually...you'll lose a lot. For me at least, risk reward is hugely important.

I would certainly agree that win rate is important.. however mine runs at around 30%. But the point about psychology being important (paramount) is always a good one.

 

[Mr. Charts]

The problem with a lower win rate albeit with a much higher average win size to average loss size, is that you are mathematically certain to have much longer runs of losing trades which is emotionally draining to all but the hardiest.
Richard

 

[HowardCohodas]

The problem with a lower win rate albeit with a much higher average win size to average loss size, is that you are mathematically certain to have much longer runs of losing trades which is emotionally draining to all but the hardiest.
Richard

I know my limitations and bearing up under a long series of losses is one of them.

 

[alexander]

The OP made some very interesting points worth further investigation. The answers cannot be easy but instead must rest on careful analysis. Some of the issues surrounding R:R and win rate are discussed in this paper, especially issues that regard trend following system. A must read:

http://www.priceactionlab.com/Literature/profitability.pdf

 

[Vaco]

I sit on the 1:1, high win rate side of the fence.

With multiple setup's occuring each day my objective is to finnish each day with a profit or an easily recoverable loss.

There is also the benefit of not worrying about missing rocket trades that would have made up for the last 2-3 losers.

This type of setup works better in choppy conditions as well inmy opinion.

 

[Jon4664]

really helpful post, I wish I had read this a few years ago!

 

[XEC10]

Copied from a post of mine on ET:

We look at a price chart and within our timescale of interest, we decide where price may go to as a possible target, and where it may go to as a possible stop, or whether price may not move much at all. We use our experience of interpreting charts, observed S/R levels, market indices, futures, news and conclude the likely future direction and extent of a price move.

Some traders have more trading experience and access to information than others, which can be used to narrow down the uncertainty of price movement. The lack of precise information means that we consider a range of possible outcomes of a trade and we try to manage the uncertainty and risk:reward by thinking in terms of probabilities, of our target being hit before the stop:


Expected_gain = Expected_win x Probabilty_of_win - Expected_loss x Probability_of_loss


We try to choose the entry, target and stop so that the expected gain is positive and that the risk:reward is sufficient:

RR = Expected_win / Expected_loss.


Each trade is unique and won't be repeated, but we take the trade on the understanding that if the trade could be repeated with our choosen entry, target and stop under the same conditions, except for those aspects that are uncertain, then the odds are in our favor the we would gain in the long run for that particular trade.

 

[buffett_man]

I actually think risk:reward is a load of b0ll0cks myself, as it implies that the market actually cares about what you're doing. Furthermore, it implies that the only strategies which can make money are those with a ratio of 2:1 or 3:1.

You set the risk, the market gives you the reward.

However I think there's a little grain of truth behind the idea of 2 or 3:1, as it clearly implies small losing trades and big winning trades.

One of the main cognitive biases leading to losing money is cutting winners too soon and running losses for far too long. If you were to stick with 2:1, that's not going to happen.

Having said all that, I trade a trend model where I don't have any profit targets.. I've been unable to find a robust mechanical system which uses profit targets, and indeed when I did trade this way, I frequently came across the dilemma whereby (e.g.) the profit target was 100 pips, and the market would get to +97 pips before reversing. Profit targeting doesn't really tell you what to do in these circumstances.

Profit target with a trailing stop

 

[alexander]

Expected_gain = Expected_win x Probabilty_of_win - Expected_loss x Probability_of_loss

We try to choose the entry, target and stop so that the expected gain is positive and that the risk:reward is sufficient:

RR = Expected_win / Expected_loss.

I agree more or less but I have a few comments to make. First, the formula you wrote holds only at the limit of large numbers, meaning that you need hundreds of trades for it to produce a stationary value for the expected gain.

Second, Expected win = expected loss = expected value of a trade. RR is usually defined as profit objective/stop loss.

Regardless all the semantics, see how this guy proves that Expected_gain > 0 is equivalent to saying that net profit is zero. The two statements are equivalent:

http://www.priceactionlab.com/Blog/2010/12/what-is-a-trading-edge-part-ii/

Meaning that if you make a net profit, expected_gain must be positive. Simplicity is a virtue.

 

[the_winner_11]

In my opinion this is totally psycological queston from this point of view.

 

[jel31]

I sit on the 1:1, high win rate side of the fence.

With multiple setup's occuring each day my objective is to finnish each day with a profit or an easily recoverable loss.

There is also the benefit of not worrying about missing rocket trades that would have made up for the last 2-3 losers.

This type of setup works better in choppy conditions as well inmy opinion.

Agreed!

I like setting orders for 1:1 then trailing stop. This means that average loss will always be less than average win and with a minimum 50% win rate, long-term the figures add up (in theory!)

Very interesting discussion!

 

[hoodoo man]

Indeed, a beginner's mistake is to assume there is a 50/50 chance of success and is predicated on the trade situation being simply random.
They also assume that there is a 50/50 chance of a 1:2 or 1:3 etc risk/reward being hit.
That is absurd, but even seasoned traders have it rooted in their minds.
Richard

Very good point. Where does this canard come from? I see some of the beginners stickies and "first steps" articles contain misleading things like this...