Risk Reward in Trading strategies

Dax - Your mate is probably scalping. Nothing wrong with that. If hes playing for ticks around the spread which is what a lot of guys on the floor did then he may not of had much room to move any stop.

I understand the statement now,. This may help:

Think of your standard bell curve that you learnt at school. This shows probabilities and deviation of outcomes. In the centre, at the hight of the curve we have the highest number of outcomes for the data set (height of people). At the extremes, where the curve is close to the axis we have the rare and unusual data. In the height example, one end we have people who are 7 ft tall, at the other we have midgets. I'm sure you get this.

Now, consider your entry price. Thats the centre point at the height of the curve. To the left you have your stop, to the right we have a profitable trade and somewhere in the rhs of the curve your planned exit. You can put your stop anywhere you like - as you can with your exit. However, you now can see that the further away from the centre point you put your stop and exit, the less chance you have of reaching that price. So if your stop is very far away, but your exit close by, you can assume your exit will be hit more often than the stop.

This is why losers move their stop. They hope that the probability of price continuing and hitting their stop is reduced. It isn't. Remember Newtons law: when in motion price remains in motion untill an equal opposite force (demand/supply) blah blah blah. Therefore, when price moves towards the stop, chances are it will continue. Thats why its better not to wait around to long for your stop to be hit. You're really hoping, against the odds. Just get out and save the pain. Sure there will be times when you get out and the trade reverses and you realise you would have been profitable.Thats life trader!

This is very crude however, as it assumes that the probability of price moving in either direction is 50-50. It wont be! But thats the skill of the trader. Assessing the probability and direction of price and planning the trade accordingly. I'm not a great maths expert as you can probably tell. Grey1 and others probably have a better grasp and understanding than I do, but I hope it helps.
 
BBB

Surely the only way of making a long term profit with a fixed 1:1 reward to risk ratio is if the short term forward distribution is likely to be skewed in your favour (lop sided not normal or bell shaped). Entry timing using TA or other means with a 1:1 ratio is effectively anticipating short term future skewness.

I suspect the reason why running the stop doesn't provide a benefit (for some securities) is because the forward distribution quickly reverts to a near normal distribution or a skewed distribution in the opposite direction once the next point of resistance or support is reached.

If there is no skewness then it is still possible to make profits using non normal distributions. Take a flatter distribution with long tails for example, then you need to use a higher reward to risk ratio to gain the benefit when the price shoots off. In these cases it is usually better to run the stop. Conversely if you have a peakier distribution than normal it makes more sense to take smaller profits since it is more likely to oscillate around a mean.
 
BBB

Very well explained, makes sense.

He wasn't a scalper, 10 tick to 30 ticks sort of trading.
 
cassiopeia, thats a good explanation. I bet you trade options?
 
BBB

Strange you should mention options. Some traders such as Nassim Taleb has made a career out of long tailed distributions. He purchases good value out of the money options. Most expire worthless, but the few wins more than make up for the losses. He knows that many are valued on limited data and take no account of potential disasters. Paradoxically, he thinks this strategy is as safe as houses over the long run. The only other thing he dares to invest in is US treasury bonds!

I used to trade options around 6+ years ago, mainly so I could profit in both bull and bear markets. Nothing very clever in terms of strategy, just a bit of TA, checking technical value and running profits.
 
cassiopeia said:
[...]
Nothing very clever in terms of strategy, just a bit of TA, checking technical value and running profits.

Cass - I think about 95% of traders think that's too simple as well...long may they continue.

Taleb (FBR) was an interesting read. I needed a couple of runs at it and put this down to his highly intellectual style (rather than my basic stupidity).

It is an interesting trading philosophy though, and I'd be really interested in data on his win/loss, drawdown etc.
 
Just caught this thread ( a few days after most of you), but for my penny's worth: 3 points

1) I agree that what works for one trader doesn't necessarily work for another. My view is that in becoming a successful trader you need a trading STRATEGY that suits your trading STYLE. This refers to the psychological aspects of trading. The style that suits you depends on the degree of risk you are willing to accept, your trading capital, etc. There are thousands of trading strategies around for each and every market, but only a few will suit your particular trading style. Find a strategy that "fits" with your trading style and you on on the right track. I believe that this is something that most beginner-traders are not informed about.

2) Bell-curves: I pondered long about why moving averages (MAs) work. When I first started to investigate trading by MAs, it often seemed to me that when I selected the "right" MA it often provided support or resistance ... like a trend line that was adpating itself to current market conditions. Eventually it hit me. A moving average shows the median (most often traded) price within a range. When the market is in a trading range then the distribution is closer to a bell curve (analogy: as prices try to pull away from the curve, the median directs the price back to the mid line, so we have a trading range). When the market starts to trend, the distribution begins to skew in that direction ... the MA therefore starts to shift ... as the price moves and traders start to follow it the momentum builds up and we have a trend underway. When the momentum slows, the median will close-in on price and the market reverts to a trading range again. What causes the median to shift significantly instead of remaining in the middle of a bell curve? Well what else moves the market ... volume. Just look what those volume spikes are doing and how the market is reacting (effort vs results)

3) Stops: yep, if your stop is too close to the median of that distribution then it will be one of the first positions in line for trading by the market makers. Keeping your stop away from the majority of the action (away from the MA) will provide you with safety net that doesn't get taken out too early.

Remember: strategy with style ... like clothing, it's what you feel most comfortable with
 
A moving average shows the median (most often traded) price within a range. When the market is in a trading range then the distribution is closer to a bell curve

I thought the simple moving average is always based on the mean unless stated differently. From memory, I think the most common number in a set is called the mode, whilst the median is the central number. The median would therefore lag more than a mean, so I agree it would be interesting to see if a shorter term median would provide more useful signals since it would filter out outlying prices more effectively.

As far as I know, most types of moving average are not very effective for predicting efficient markets since the market jumps too fast, that is once you have got aboard the trend changes back. I was reading a book called 'smarter trading by Perry Kaufman'. He shows that highly developed markets such as the Dow actually buck the trend (on the scale of 5 days+), in other words you can profit from going short when the trend turns up and vice versa! I doubt if it is worth it after charges through.

when I selected the "right" MA

That is the problem which one do you use? Kaufman uses something called an adapted moving average which effectively varies the trend speed in accordance to how much noise there is in the price to avoid whipsaw losses. Seems sensible!
 
cassiopeia said:
From memory, I think the most common number in a set is called the mode, whilst the median is the central number. The median would therefore lag more than a mean,

Not necessarily.

The mean is (N1+N2+N3...+Nm)/m

The median is the N(m/2)th (sorted ) value in the series.

If you had an unbalanced sequence (lots or very few - very small/very large values) this could skew the mean and result in the mean lagging the median.
 
Hi there

I always trade for a 3:1 minimum on Flags using using MACD to confirm break up. If the breakout uptrend is comprimised before profit target I sell half my position and if then MACD breaksdown in sympathy I sell the other half.

The other indicator I use is H&S, as placing a stop above the second shoulder usualy allows for a 3:1+ - P:L ratio. In this situation I set an OCO and go to town for the afternoon.

I find this strategy very profitable and psychologicaly simple to implement.

I have also run successful strategies using a permanent 1:1max ratio using a trail stop which averaged 20 pips per day for the 10 months I back tested, and traded it . However it was less profitable, more time consuming and more stressful than what I have now.

I am not suggesting that a strategy using a larger P/L ratio necessarily more successful although I do know, from experience, that when combined with specific technical indicators a 3:1+ ratio significantly approaches the success rate of a 2:1 or 1:1 target when trading the same breakout.
 
Position sizing software

I'm looking for position sizing software that can handle at least 5 portfolios. Any thoughts or experience to share would be appreciated.
 
cassiopeia said:
BBB

Surely the only way of making a long term profit with a fixed 1:1 reward to risk ratio is if the short term forward distribution is likely to be skewed in your favour (lop sided not normal or bell shaped). Entry timing using TA or other means with a 1:1 ratio is effectively anticipating short term future skewness.

I suspect the reason why running the stop doesn't provide a benefit (for some securities) is because the forward distribution quickly reverts to a near normal distribution or a skewed distribution in the opposite direction once the next point of resistance or support is reached.


Both parts of this make a lot of sense to me. What I am looking for in shortish term trading is a factor that skews the distribution. I then speculate how long the forward distribution will remain skewed. The 1:1 idea (and its close relative, the fixed number of ticks objective) fit into this pattern by attempting to exploit the skew while it lasts. So, in 5 minute bars, it is probably not sensible to expect a skew to last more than 4 or 5 steps; not really enough to generate a very robust value for the option. If you go for more distant horizons, and for higher risk/reward ratios, you run into the likelihood that the distribution is normal.

But in any case, the idea behind strategies that posit high ratios of reward/risk and accept lower success ratios is one with which I instinctively feel very ill at ease; the ex ante odds on a successful outcome after an individual act of clicking the mouse are by assumption low, and evidence that the strategy is sound - if it is - will accumulate in a lumpy manner, with frequent drawdowns. And that is if there really is an edge in favour of the strategy.
 
Top