It's YOUR Choice

BBB

Experienced member
Messages
1,071
Likes
3
Here are some methods of managing your money:

A/ A fixed percentage (say 2%) on ALL trades

B/ Start at say 1.5%, after 2-3 consequtive winnining trades, increase by 0.5%. Decrease by 0.5% after 2-3 consequtive losing trades. This allows you to adjust your size for when you are 'in gear' with the market (quick n dirty optimal f type approach)

C/ Risk 1% if the entry signal is only in the intermediate time frame. 1.5% If only in your longer term time frame, 2% if it exists in both time frames.

D/ The nightingale system: If you lose, double you position size on the next trade - hey this losing streak must end soon and I'll make it big when it does!

E/ Pah! MM ain't so important. Trade entry technique is the secret to trading so I'll continue to add to my long position each time a bar makes a new high untill I'm just shy of a margin call!

Which do you prefer?

Do you have any other ideas?

Who thinks D will apeal more to SB traders?

:LOL:
 
My personal approach is A.

Except i'd probably go for a slightly more agressive 3% risk.

B and C both have merit, but I personally think that simplicity works best in the markets.

D is for short-term gamblers, not people who want to make a living.

E is false - entry comes 3rd in order of importance after position sizing / money management and state-of-mind.


This of course is all just my opinion - feel free to rip it apart.
 
Currently use A- fixed percentage.

PS I am a spreadbetter but D is certainly very unappealing to me - I think you mean the martingale method?
 
Yea - sorry you're right it's martingale.

True, simplicity works best.

My preference is C because if I have agreement in multiple time frames, then I can assume that the trade has a higher probability of moving my way - so I can afford a larger size than if I am not as sure.

The key is that I dont think any answers are 'correct', the key is to find what you are happy with. Just like the amount of risk you assume.

Believe it or not, friends of mine who traded futures when the LIFFE floor was open traded with E! But remember - they were closer to the action so could bail at the first sign of trouble. Just goes to show that MM is intertwined with the method you are trading.
 
I trade mostly NASDAQ stocks at the mo' but I am considering going back to derivatives (futures).

As for time frame, I dont really have one! I trade what the chart tells me in what ever time frame there is least risk. Most of my analysis is done on the daily chart. If the weekly chart supports the view, I may be holding the position for a few days. If the weekly doesn't support it, but the signal is still strong on the daily chart, then it may be an intraday trade (the daily chart is now my long term time frame).

I will manage the trade on a 15 or 30 min chart during the day. If that trend falters then I take my money and run.

Take today for example. AAPL was looking fine on the daily (the tail /intraday reversal) caught my eye with the stock being in the early stages of an uptrend. Iy could have been good to hold for a few days. The weekly chart wasnt so strong so if I took the trade I'd be looking at an intraday trade with an entry of 21.18. As there is an interest rate announcement at 2:15 EST I figured today wouldnt be so good for an intra day position due to the volatility that it could cause. I am staying out today!

Hope this helps.
 
Simple $$$$$!

I won't do it though until there is more liquidity on the electronic markets.

I refuse to trade where orders are executed on the floor. This includes NYSE. I'm an ex local and have seen with my own eyes the corruption and theft that goes on to customers accounts. Poor customer just thinks its a bad fill and doesn't even know!
 
BBB,
Martingale method ( increase pos size afrer a loss, decrease after a win ) and equity curve method ( increase size when equity curve falls below its MA or increase size when you cross above the MA ) DONOT WORK

Reason :-- they assume dependency between trades

Fixed Fractional :-- The problem with this method is risk of each trade is the same, regardless of the number or stocks/ contracts. Also what is the most optimum percentage ...

You can use a ad hoc level of 2% or an Optimal value for each trade ( look at Ralph Vince Portfolio managment Formula or Secure level (look at Leo Zamansky 's Work in TASC ) or Monet carlo simulation by bryant in TASC Feb Issue 2001

There are some other Excellent position sizing techniques you have not mentioned such as margin Plus and Fixed ratio pos sizing

I am using Monecarlo simulation method in my scanner to asses the risk and hence adjust the pos size to make trade less risky compare to historical perferomance of underlying stock/instrument .. Look at the dynamic risk adjusted trade on MR NAZ and CHART Thread for the result..


If you are into academic risk/pos sizing techniques let me know
 
Last edited:
look at the MEDI today .. The correct pos sizing depends on the risk levels which dynamicaly is adjusted bar by bar.. Once the risk is determined then the pos size can be choosen to make the trade worth while on the time frame you are trading .. In my case be a 1 minute..

If I was to take the trade as it is now. I would have to have tiny pos size because the risk of going long is 66% with 18C to gain against a loss of 35C..

I would be better off to bet on the black or red in casino in this case would not I ...? Sure I would
 

Attachments

  • medi.jpg
    medi.jpg
    77.4 KB · Views: 677
I would be very interested in talking to you about the position sizing techniques you mention.

The only techniques I really have experience of are the percentage of equity and the percentage volatility techniques.
 
Grey 1
I have pm'd for some info but could you explain why equity based position sizing doesnt work - I just don't understand what you mean about dependency between trades (bear of <0 brain)
Thanks
Ron
 
Thanks Grey1.

I agree - each trade is its own event which is why I'm not such a big fan of the first two.

I use my third example which seems to make sense to me.

What are your thoughts on the RAlph Vince book? I've read some good reviews about it. Some of the ideas I've came across like optimal-f seem to take the concept of a trade as a sequence of random events. This is fine if you trade a mechanical system/use a program to get you in and out of trades aka Thomas Stridsman, but I dont. Thats perhaps why I've stayed away from complex models.

Am I doing myself an injustice then??
 
Damian -

Good job I didnt go long AAPL today eh!

Grey1 -

Thats some impressive software you have there.
 
Rognvald

Get a bag and put three balls in it.. 2 Whites and a red .. now draw one ball out and lets say it is a white.. . Don’t put the ball back and draw another ball out , lets say the next ball is again a white.. Well the chance of the last ball being the red is 100%.. This means by eliminating the first two whites from the bag we have increased our chance of having the red ball .. This is called dependency …
Lets look at the same scenario in trading .. You lose the first trade, then you lose another trade,, doe s it mean now it is your turn to win ? NO . this is called independency …

THE ODDS OF NEXT TRADE BEING A WIN IS NOT RELATD TO THE LAST TRADE..


Equity curve method assumes dependency .. This is why it does not work ..


BBB,

Ralph’s technique ( OPTIMAL F ) does not take the drawdown into account.. Theoretically is a sound technique but not practical for trading .. SECURE F Same as above plus that it brings Drawdown into the equation .. GREAT .. Just one catch .. Drawdown is calculated based on Historical sequences of Trades .. ( Not a big catch really as there is little choice in assessing draw down Except relying on history )


PS:-
Skilled traders look into correct entry first.. It is crucial to enter a trade where the price immediately moves in your favour than spending hours and hours to optimise the position size/ stop loss level and the exit and do all kind of curve fittings to have a profitable strategy.. ENRTY is the key for success full trading if you are a trader .. However if you are a mathematician then choose a random entry and optimise the exit to have an attractive Return on your account..
 
Grey1

Sure entry is important, I just dont think it merits the attention most give it. I personally think self discipline in perhaps the most important aspect. Either way, a trader who knows how to manage his trades, eliminate risk as much as possible and of course manage himself should be able to be profitable by entering on the flip of a coin. I'm not claiming I could do this by the way ! ( But I'm sure it would be close/break even.)

I also agree with not curve fitting your trades. You never know whats around the corner.

Where can I find out more about the montecarlo system you use? Can you calculate the parameters by hand/in your head or do you need a software package?

As we all agree each trade is its own event, surely past results shouldn't enter our equation?
 
Grey1,

Couple of things I don't understand:

a) You are using a monte-carlo simulation. As this is simply a random event generator, based on historic events, why not simply utilise the full stats for the time period(s) you are examining ? i.e. 5 day range=avge 2% with 3% SD.

b) The type of risk analysis you are referring to will always spit out odds based on regression to the mean - which effectively means you are swing trading (albeit under a different name) and would therefore not get into some long and protracted trends.

Sorry if my understanding of your comments is incorrect, but seems pretty clear cut to me.

Rgds,

The General.
 
Grey1 said:
THE ODDS OF NEXT TRADE BEING A WIN IS NOT RELATD TO THE LAST TRADE..

Does everyone here really believe that?

I can see the view being formed from a very pure abstract analytic view of the markets, but I kind of doubt that its entirely true in practice.

Grey1, you yourself advocate "2nd pullback". Now isn't that implicitly saying that the previous action carries forward into the future? Aren't you really saying that if you traded every pullback the second trade after a previous failure would have more chance of winning than the first?

Not necessarily the most practical thing in the world and I'm not really sure if it is something that can give or take away an edge. But for example, lets say you have a robust breakout strategy that statistically given 90% winners over the past 10 years. Now treating that as a straight pure probability matter there would be a 0.1^100 chance of getting 100 consecutive failed breakouts wouldn't there? But at a gut level I just don't think that the market is a pure mathematical construct and I'm not really sure I believe as a fact that you could get 100 consecutive failed breakouts on a robust system.

It comes down more to gut feel and instinct than anything you could ever prove, but whilst I think statistical analysis is good and in fact necessary, I'm not sure that the markets are a pure construct that can be analysed properly that way (although I doubt there is anything better and I would have thought the problems occurred more at the tails).

wysi
 
BBB

Quote “Where can I find out more about the montecarlo system you use? Can you calculate the parameters by hand/in your head or do you need a software package? “

I have given a reference in my first post..Bryant work is a good start .. you wont be able to calculate all these in real time .. you need a software package to do that for you.. you as a trader should just concentrate on the correct execution

Quote “As we all agree each trade is its own event, surely past results shouldn't enter our equation? “

Not quite agreed on that . if we did then we would not be on this BB.. market is a semi efficient entity and our job would be to exploit the packets of inefficiencies in every day trading.. Our job would be to look for these packets and trade them .. This is why we sit down and analyze the charts.. However generally speaking the trades are independent ..

General :-

Quote “a) You are using a monte-carlo simulation. As this is simply a random event generator, based on historic events, why not simply utilise the full stats for the time period(s) you are examining ? i.e. 5 day range=avge 2% with 3% SD. “
Monte carlo is a random number generator, .. “ yes, this is why I am interested in it.. I can use the result of the simulation and replace it with its probability distribution to have the best Fixed fraction %.. This way I have replaced the “trade “ with a “ distribution of Trades “.. So what I have effectively done is to randomize the sequence of trades and for each sequence have a return and a MAX drawdown .. The next step then would be to find a confidence interval for both return and drawdown and find the “F “ to maximize the return .. if you have a look at the MEDI trade the deviation from the mean is well more than 2SD and the price kept falling and in fact if you did take the trade at around 2SD you would have lost the trade… This is why the mean reversion must be coupled with risk analysis to work properly other wise it wont.. An example of it was yesterday trade of AMZN when paul ( trader333 ) got a long signal well before me and I had to wait a bit longer to call the trade.. ( the trade was in real time )..

Quote “The type of risk analysis you are referring to will always spit out odds based on regression to the mean - which effectively means you are swing trading (albeit under a different name) and would therefore not get into some long and protracted trends.”
SWING TRADING ? not at all .. When stocks/instruments fall they start going into trend in lower time frame .. This is where trend is born .. one must get into the trend well before other notice it and sell it to those who are wanting to ride the middle part of the trend…. Sell into strength .. We sell to those who are sitting for their MA to cross..

Have you ever thought that there is not such a thing as TREND and all you see is swing but in higher time frames .. Heyyyy..


Hope this helps..
 
Top