More trades means more risks ?

[MikeW]

A risk profile can also be helpful to highlight expectancy & small %possible large negative outcomes which you need to deal with emotionally (& that's assuming your assessment of outcome are reasonable):
e.g.
Expected Outcome % Probability
+20 5
+10 10
+5 80
0 3
-10 1.5
-100 0.5
total +5.35 100%

EMV= (20x5%)+(10x10%)+(5x80%)+0-(10x1.5%)-(100x0.5%)

BUT you still need to ask yourself if you're comfortable about the v. low assessed % risk bad outcome. Obviously these figures are made up for illustration purposes but hopefully you get the drift.

(not sure when I post the relevant %'s do not appear beside the outcomes just read from top down i.e. +20=5% & so on)

Regs.
Mike

 

[Glenn]

Gre1 & Paul
I understand expectancy and it's formulae - there are others too. Optimising expectancy is important - paramount even.

However I think that we are on different tracks here. You are talking about the expectancy from each trade (i.e. what the return is expected to be, compared to the amount risked).
What I am talking about is the probability of being able to repeat that in a number of short term trades compared to fewer long term trades when they both have the same expectancy per trade.
(Incidentally, I'm not taking a particular side in this debate, just exploring the numbers and keeping an open mind)
So going to the formula from Grey1.
"Expectancy is .. ( Average win* probability of win - Average loss* probabilty of loss ) "
Using my numbers from before -
The short term trade expectancy is:-
(10 x 60% - 2 x 40%) = 6-0.8) = 5.2 i.e. for every £2 we risk, we stand to gain £5.20
For the long term trade the expectancy is :-
(50x60% - 10x 20%) = (30-4) = 26 i.e. for every £10 we risk we stand to gain £26.

So the two methods have the same expectancy (2.6 times the amount risked). This means that historically on average over a long period, both systems have returned £26 for every £10 risked.

Now let's look at the probability of the 'expected' return happening over a number of future trades.
Take the short term figures. For every £2 we risk, we stand to gain £5.20.
The chances of that happening are on each trade are 60% (the probability of a win).
Over a number of trades the expectancy per trade stays the same.
The chances of achieving that expectancy outcome on every one of 10 trades in a row and achieving 100 points profit is 0.36%.
With the longer term system, with the same expectancy per trade, the chances of the two trades producing 100 points is 36%

On that basis, the larger trades seem the best option.

In addition, the variables which the maths doesn't account for are:-
1. The commissions, as Grey1 has already said.
2. The human side - The skill, discipline and attention of the trader. If it is harder to correctly manage a larger number of smaller trades, then there is more risk.
3. Flat trades.

Glenn

 

[Grey1]

Glen,

Thank you for your initial post and your further replies .. T2W is the place to be for coaching and educating our traders..

Few points..

1) Optimising expectancy is different thing to your original question .. Your initial post was more Trade means more risk and I disagreed. and gave you the reasons ..

2) Optimising the expectancy score can be done through Money management to take advantage of the edge , which has been discussed in here before by myself and others..

3) The average win is a net gain after the slippage.

4) a system with a negative expectancy such as roulettes can pay off in short term ( Luck ) but will beat the punters in long term no matter what .. So you wont be able to optimise the roulette system to your favour by Money management ..


Lets get back to the original question .. DOES MORE TRADE MEANS MORE RISK ... Answer is NO .. It means Over trading that is all..


PS:-- Just realised did not address the probability of the expected return you meantioned in a short term ... Any thing is possible in short term including getting 4 heads in a row to give us a 100% false win rate it is called LUCK .. Now If i ask you the probability of getting 4 rows you would tell me is (1/2)Factorical 4 which is 1/16 and not 100% only because you just had 4 heads in the row.

 

[wysinawyg]

A quiz for our Traders...

If I had a system with negative expectancy would an excellent money managment system make me a winner ?

Depends exactly on the system, and your Vegas example gives the possibility when it can be true.

Blackjack is undoubtedly a negative expectancy for the punter. Even counting cards I think the game remains a negative expectancy taken as a whole (on the basis that you can't decide which hands to play without people getting suspicious). However by betting larger on the hands where you have the edge you can turn an overall negative expectancy game into a positive one, purely through the money management.

wysi

 

[BBB]

wysi - Isn't that a beutiful thing! The flexibility of thought and application.

 

[Glenn]

Grey1
I understand what you say, and have to say I'm not convinced by your line of thinking.

In this statement you appear to contradict yourself:-
"DOES MORE TRADE MEANS MORE RISK ... Answer is NO .. It means Over trading that is all.."
By overtrading do you mean trading too often, so that you spoil your rate of success ? If so, then you appear to suggest that more trades is a bad thing.

Looking at a long term situation (as long as you like), there will be 5 times as many small trades as big ones.
For every 100 larger trades you need to do 500 smaller ones to get the same profit. That's a lot more to get right.
It may be that you personally are quite capable of correctly managing so many trades. But would it nonetheless not be easier and therefore less risky to do it with less ?

Glenn

 

[Grey1]

Glenn,

By over trading I donot mean spoiling your rate of success.. There are scalpers who trade 200 times a day and scalp 10C in every trade.. Since they have a postive expectancy system then their over trading or if you like trading 100's of times aday just generates more profit..

Quote "
For every 100 larger trades you need to do 500 smaller ones to get the same profit. That's a lot more to get right.

A system which trades 100 large trade could have a negative or lesser expectancy score than that of a system which trade 500 a day

A trader could trade 1 millions times a day and still be a winner ( I know this is far fetched ) because his system is profitable or has a Positive expectancy

Another trader could trade just 20 times and be an over all loser for the day , simply because he has no edge or he his system has a negative expectancy

Hope I am clearer this time..

 

[Glenn]

Grey1
"Hope I am clearer this time.." Erm, no - you appear to be moving the goalposts
My earlier comparison was between two systems with the same expectancy. If they have different expectancies then it's a lot more complicated to make a comparison.

If we stay with two systems which have the same expectancy but different timescale/frequency, then I think we can reach agreement eventually

Glenn

 

[Grey1]

Glen.

Your post was "more trade was more riskier" and I replied to that based on simple expectancy formula which we both agree on it.. The formula does not have ""frequency " as a variable in it .. so your statement was in accurate..

You then said about optimising expectancy and I replied that is Money managments job.

Now you are telling me what if we have two system with similar expectancy which one would trade 500 times and the other 100 times the first system is riskier..

Have I got you right ? if yes then can you explain why

regards

 

[Glenn]

Hi Grey1
Quote "Your post was "more trade was more riskier" and I replied to that based on simple expectancy formula which we both agree on it.. The formula does not have ""frequency " as a variable in it .. so your statement was in accurate.. "

Whilst I agree the principle of expectancy, I don't think that it is the answer to this question.

Expectancy just tells you whether you have a method which will produce a profit over time.
Just because a system has a positive expectancy, that is only theoretical. The trader still has to correctly manage every trade in order to achieve it in reality. Question is, is that harder with more frequent, smaller trades ?

What I am trying to do is compare two systems which have the same expectancy but different time periods and frequencies, and ask the question 'Is one riskier than the other ?'

The risk comparison can be anecdotal or mathematical - whatever people want to discuss. (Although there aren't many of us discussing it )

I mentioned optimisation of expectancy only in passing, just to show that I am aware of expectancy and how it works, not to bring it into the discussion. My mistake.

Quote "Now you are telling me what if we have two system with similar expectancy which one would trade 500 times and the other 100 times the first system is riskier."

Well, not telling you actually, just suggesting it as a possibility, which is the point of the thread. I have explained why one method appears to be be riskier, using some maths and anecdotal thoughts.
I think that you disagree with the anecdotal thoughts - fair enough. We can agree to differ on those.
As for the maths (probabilities) I don't think that you have made a case yet.

Hope that explains.
Regards
Glenn

 

[wysinawyg]

Glenn,

I see where you're coming from, but it only applies if you think someone is less able to perform the trades at speed and you're basically talking human error.

i.e. two strategies both of which "should" give 500 points over a year, the first of which does one trade and the second does 500 trades. Clearly on the first system a bit of human error is unlikely to make a huge difference to the result, but on the second system you could pretty much guarantee that in practice the person would substantially underperform the theoretical.

I agree with that to an extent, although it becomes less of a factor when you are looking at reasonable sized moves as opposed to small scalps.

The flip side that grey1 is getting at is that if you expect average profits of 500 from say sytem 1 which makes one trade a year or system 2 which makes 10 trades a year, then system 2 is much more likely to give you close to 500 points than system 1 when taking any particular year. It isn't that system 1 will do less well on average (and in this case the human error part is likely to be much less of a factor) just that in any particular year you will have much more volatile results. As such system 2 can be seen as "less risky" because your chance of getting average results (rather than catastrophically bad results) is much lower.

wysi

 

[Grey1]

Glen,

I had a mathematical view towards your statement this is why I said It was in correct.. Human error, and stuff like that was not part of the equation and I leave that to individual traders..


We must have a solid and logical mathematical view toward our strategy and once that is acheived then train ourself and our mind to stick to our rules/money management other wise we should not be trading..



regards

 

[DaveGos]

Glenn
"Each time you take a new trade, you take on a completely new risk.
i.e. More trades = More Risks (and higher trading costs)."

As a statement of fact that's unarguable, but providing your system is net positive then More trades = More profits, in fact the more trades you do the closer you should come to the systems winning probability, i.e with a system providing a winning probability of 50% it is quite possible over say 100 trades to produce only 40 winners but far more likely over a 1000 trades to produce 500 winners, thats Mathematics.
It may be that I am taking to simple an approach but it seems to me that fewer trades could lead to increased risk because the population of trades is not large enough statistically to deliver the number of wins that systems win probability calls for.
Of course if your arguing that more trades increases the chance of human error then thats another matter.

 

[Glenn]

Grey1
Quote "I had a mathematical view towards your statement this is why I said It was in correct.. "
But the maths you have mentioned was about expectancy. We have eliminated expectancy by suggesting that the two systems under comparison have the same expectancy.
Quote "...We must have a solid and logical mathematical view toward our strategy"
Agreed. Do you have a probability (maths) case to offer in respect of the two equal-expectancy systems ?

Wysinawyg
Quote "I see where you're coming from, but it only applies if you think someone is less able to perform the trades at speed and you're basically talking human error."
OK, well that seems like a rationale in support of the argument that more is riskier.
If you have more opportunities to make an error, then maybe you will. Perhaps an argument which suggests that the less experienced you are, the less often you should trade ?

However does it not also apply because of the probability (maths) argument ?

Quote "i.e. two strategies both of which "should" give 500 points over a year, the first of which does one trade and the second does 500 trades. Clearly on the first system a bit of human error is unlikely to make a huge difference to the result, but on the second system you could pretty much guarantee that in practice the person would substantially underperform the theoretical."
So again, something in support of the argument.

Quote "I agree with that to an extent, although it becomes less of a factor when you are looking at reasonable sized moves as opposed to small scalps."
Sorry, not quite with you here. Why would that be ?

Quote "The flip side that grey1 is getting at is that if you expect average profits of 500 from say sytem 1 which makes one trade a year or system 2 which makes 10 trades a year, then system 2 is much more likely to give you close to 500 points than system 1 when taking any particular year. "
OK, well that's changing the parameters a bit. One trade per year is not really trading is it ?
Nevertheless, the maths don't appear to agree with you.
Using my earlier trade numbers:-
Longer term trade - Chances of a 100 point win from one trade is 60%
Short term trade - Chances of 100 points from 10 trades is (60%x60% .....x60%) = 0.36%

Quote "It isn't that system 1 will do less well on average (and in this case the human error part is likely to be much less of a factor) just that in any particular year you will have much more volatile results. "
Quite likely for that example, but not a good sample size really (imo).

Quote "As such system 2 can be seen as "less risky" because your chance of getting average results (rather than catastrophically bad results) is much lower."
Erm - not quite with you again. If your chances of getting average results is lower then isn't that a bad thing ?
As regards catastrophically bad results, presumably from a big hit (?), then it's hard to compare the two. When you are in a trade, however long or short, this could happen to you. It could however be argued that longer term trades spend longer 'in the market' and are therefore more likely to suffer a big hit.

DaveGos
Quote "it seems to me that fewer trades could lead to increased risk because the population of trades is not large enough statistically to deliver the number of wins that systems win probability calls for."
Agreed. I was trying to use examples which could be classed as trading rather than buy-and-hold, so that there was enough opportunity to trade both.

Quote "if your arguing that more trades increases the chance of human error then thats another matter."
OK, so do you think that's true ?

Glenn

 

[DaveGos]

Quote "if your arguing that more trades increases the chance of human error then thats another matter."
OK, so do you think that's true ?

Glenn

Yes I do, some years ago I was involved in designing tests aimed at measuring a specific aspect of cognitive function. The test involved a simple repetetive task involving identifying specific instances of a coloured letter in a stream of rapidly changing coloured letters. We found that a statistically significant problem was that the longer the individual performed the task in one sitting the greater the increase in errors as the task progressed.
What they were doing was anticipating the target letter because they believed they were seeing sequences, when in fact the numbers were generated randomly.

This is why as traders if we are not aware of this tendancy to trade against our own strategy and ensure we have the self dicipline this requires we will fail.
It also helps to take a break every now and then, have a cup of tea, walk the dog, talk with your wife.

Dave

 

[Glenn]

Dave
Interesting. Certainly must have a bearing on the overtrading question.
Those traders here and elsewhere who trade very often must be able to keep their concentration well. Perhaps one or two of them will speak up and let us know how they do it ? They may have some personal methods to share, or just be lucky robots
My limit seems to be a couple of hours when day-trading. I have come to prefer to trade over longer periods (days), especially if the returns are the same. But then I'm not as young as I used to be.
Glenn

 

[DaveGos]

Glenn

I've never traded side by side with others but from my experience at work, I'm a project manager, the older you get the more risk aware you become. So heres hoping our trading improves with age.
Dave

 

[Grey1]

Glen,

I am not going to bring the human parameter into this equation .. That depends on individual characters and disciplines ..

Your statement more trades means more risk is mathematically wrong..

This is the case I am putting forward

1) In the formula number of trades is not a variable.. hence irrelevant to talk about trade numbers

2) if you make 10 Trade or 10 million trade the expectancy formula only asks you for Average win/loss during the trading period. ..and not how many trade one made to get the average win

Quote " However does it not also apply because of the probability (math) argument ? "

Probability distribution needs a run of the trades over a large period.. I have said that before , if you throw a coin and get 4 heads probability would tell you the chance of this happening is 1/16.. based on such a small sample (only 4) one must not risk his capital. In fact it is point less to talk about probability of any thing over a short term period SO your argument based on probability of shorter sample is flawed.


I think I have explained and expanded enough on expectancy..

Conclusion for our younger traders :--

1) More trade does not make more risk .. it means more reward or more loss depending on your system expectancy.. Simple as that.

2) System with positive expectancy create more profit if traded more (in long term ) in short term any thing is possible.. Even casino's might lose in short term ..

3) Money management is the key to create wealth ONLY and ONLY in systems with POSITIVE EXPECTANCY .. NO ONE CAN MAKE £££ using MONEY MANAGEMNT Techniques if his trading system has negative expectancy.

The above conclusion is pure math and has been around for 100's of years and are not negotiable .. LAS VEGAS WAS BUILT BASED UPON THE ABOVE RULES..



Human error, emotions, physical, mental tiredness, dog jumping out of the window while trading and stuff like that does not interest me. All I am interested is if the foundation of my trading strategy is solid as a math then I can carry on to teach myself discipline to enforce the rules ..

I would expect our trader study their trading system and make sure , they have one with positive expectancy ..


PS:-- I have tried not to be distracted from the core of argument because I feel this is fundamental for your younger traders to understand..

 

[wysinawyg]

Glenn,

<i>Using my earlier trade numbers:-
Longer term trade - Chances of a 100 point win from one trade is 60%
Short term trade - Chances of 100 points from 10 trades is (60%x60% .....x60%) = 0.36%</i>

Are you talking about just binary outcomes? In which case with the Longer trade your average is 60 points but the actual result will be 0 or 100. Now if that is your living (or even your pension fund) an outcome that stands a 40% chance of giving you 0 is risky, really risky. Over the course of 20 time periods you should average 60 points but you'll have lots of very bad time periods.

The short term trades on the other hand may only have 0.6% chance of getting 100% (I think that is right rather than your 0.36%, but its pretty irrelevant) but they only have a 0.01% chance of you ending up with nothing. So whilst you're unlikely to get 100 points you stand vastly less chance of coming out of the time period with 0 points and you should come fairly close to the average 60 points.

Now to me if I cared about how much I got I would much rather have a good chance of getting 60 in every time period and very little of getting nothing rather than what is nearly a toin coss between 100 and 0. The difference in volatility makes a huge difference in the real world, although the expectancy is the same one pretty much guarantees an income of some sort for each time period whereas the other pretty much guarantees no income a lot of the time (and both will give you the same on average).

On the maths argument I think you are just plain wrong and favouring a winner takes all gamble against an equivalent steady income, which seems much more risky to me even where the average is the same.

From a more human mechanics point though there is an element of truth to it. As I said you could argue that human error may well cost you a point a time which would be a big difference with 10x10 trades and much less with 1x100. On the other hand though if you miss the 1x100 through human error you get nothing at all whereas if you miss one of the 10x10s you've only dropped an average 6 points and will still probably come out with an income.

wysi

 

[TheBramble]

I'm probably missing something but...

If all your trades have the same probability of success it doesn't matter whether you run 1, 10, 100 or 100,000 trades. Your risk is exactly the same in each trade AND overall.

If you're saying "what's the probability of this trade failing OR this trade failing" that's 0.4 + 0.4 = 0.8.

If you're saying "what's the probability of this trade failing AND this trade failing" that's 0.4 X 0.4 = 0.16.

But on an individual trade-by-trade basis, it doesn't matter how many trades you run. If they all have the same probability of success (and your running a MM system that doesn't skew the data) you've got the same level of risk.