Improving your trading strategy using probability

Originally posted by Fluid
Are you seriously using this P2 kid as an example for position sizing/risk management? Matter of time, just a matter of time...
More...
As an example, yes. As a thorough proof of validity of such methods, no.

P2 is well-respected on ET and has posted many times on his trading methods. As with all strategies, the devil is in the details. I am curious as to his risk management methods for adding to losers. Hence the invitation for P2 to post on this thread.

And considering P2 is about to manage a fund using his same style of trading, perhaps he's done some diligence on risk management specific to his methods and can offer some intelligent discussion to this thread.

I've also given a few examples of using variable position sizing/trade selection, including my own specific use of portfolio rotation.
 
Originally posted by darkhorse
for some reason I can hear the guys at LTCM saying this in soothing tones to their clients, circa early 1998....
More...
The failures of LTCM have been thoroughly dissected in various texts. The failure of LTCM is not proof that variable position sizing while using stat arb strategies is not an effective strategy.

With proper risk controls in place LTCM failure would not have happened. Their profits would not have been as high either. Did I mention one has to look at overall return-on-risk? LTCM's failure was due to a much more fundamental reason...

Also, stat arbs are not the only strategies where variable position size/trade selection techniques are used. I know a lot of seasoned traders look down on such an approach. I am simply saying it is a valid method to increase return-on-risk, albeit with the proper approach and precautions. My long-term trend following portfolio rotation strategy is an example. In fact, I consulted on its use to a large hedge fund that uses the very same concept to increase return-on-risk with less diversification.
 
Actually, the precious flip of a coin can influence the next flip. If you want to be super-critical about cause and effect, you could argue that merely flipping the coin will cause it to wind up in a specific position that it otherwise would not have ended up in if you didn't flip it to begin with.

Nothing takes place in this universe without a cause. All effects can be linked to something else.

There was someone who once suggested they would fade the 51'st flip of a coin if the previous 50 flips came up all heads.

This is how the world works:

A mathematician would insist that the 51'st flip was truly random.

A realist would argue the coin has two heads.

Do you want to fade a two-headed coin?

Nothing is "random" in the universe. Anything which appears significantly random is probably operating off some level of chaos. Quantum mechanics don't count.
 
Originally posted by dottom

The failures of LTCM have been thoroughly dissected in various texts. The failure of LTCM is not proof that variable position sizing while using stat arb strategies is not an effective strategy.
More...



'twas (mostly) a joke... not really disagreeing w/ you dottom, except to point out that the danger of outliers is tough to quantify... i think my playful side was mainly responding to the extremely high confidence in your tone re backtesting as a safety measure...
 
Interesting topic.

I have a friend who uses a breakout/reversal strategy that waits till the model has failed 2 times then takes the next trade. Short term (hourly bars) on the Euro Stock Indexes. It has worked well for him for a number of years.

Also I'm a big believer on 'The Law Of Diminishing Returns'. Funny thing is that I don't know if it actually exists! What it basically states is that as 'anything' gets older its life expectancy gets shorted. You, me, this board, the screen you're looking at, Earth etc, everything.

Whether it exists or not it makes a lot of sense to me and I use it often in my life (not necessarily for trading).

Another interesting point is this, the world record for the same colour coming up in a row on Roulette is about 48 in a Puerto Rican Casino in the 1950's, now what nobody can ever explain to me is that how come something like this (even taking the 0 or 00's out) can happen. Surely a run of at least 100 is possible?

These are just some thoughts, please all you scientists out there, don't attack me :)
 
Originally posted by AAAintheBeltway
dottom,

Good stuff. What are you using now?
More...
20% of my equity is in NN EOD model. Average position is 5 days. 80% of my equity is in day-trading where I use a stat model of price/volume/momentum patterns. Momentum is not the traditional ROC, but is taken from DSP in frequency domain. It acts only as a filter to indicate "windows of opportunity" where the stat model has higher % chance of success.

I also have a couple other "setups" that I integrate into my day trading system. For example, I objectively defined Dinapoli's "railroad tracks" setup and added a volatility filter. (hey, there was an RR entry today!)

I used to trade long-term trend following similar to Turtle method but with different entry/exit criteria to minimize whipsaws and also some basic pattern recognition which identified sup/res to use more meaningful stop placements. (e.g. the Turtle method would have you enter on a 20-bar breakout... but suppose the 24-bar breakout was also a 60-bar high but 20-bar breakout was only 20-bar high, wouldn't the 60-bar breakout be more significant entry point? Similar checks for stop-losses.) Think of Turtle + Aberration + basic sup/res pattern recognition = my long-term trend following method.

But long-term trend following can't beat the level of consistency I'm getting from day trading, which means not only higher returns, but less risk, smoother equity curve, and ability to compound money faster. After talking to some traders at prop firms who were consistently making $1-2k per day with only a few losing days per month, I firmly believed that day trading was the way to crank the wheel of positive expectation.

So I started with some a priori theories + empirical observations -> rigorous backtesting (taking all necessary precautions... I used the same disciplines from NN analysis in backtesting my a priori system) -> forward trading. I started relatively small and have been adding in conservative amounts. For example, I am only trading 1 ES contract per $10k equity when I could easily have done 1 ES per $3k and never even flinched at max drawdown. Of course, as your contract size gets larger you start becoming more conservative, less willing to accept the volatility regardless of what the risk metrics say historically was acceptable. Diversification helps to alleviate some risk.

I still have my long-term trading model and portfolio rotation. If/when I retire I'll just trade my long-term model and take life easy. Of course, once a day trader always a day trader. I just can't imagine how boring life would be to sit on a beach and check your positions just once a day. Or worse, program it via broker-assist and get a daily update.

I hope to eventually run a fund (I ran one for 3yrs via power of attorney for private investor using my long-term method, but I wasn't a CTA or anything. Just a handshake and POA. He passed away. His wife didn't know about his account, and was quite happily surprised when I showed up.) I'll invest large portion of my own capital in it. I'd like to build a 2yr day trading record first. I do have my long-term record already. But I have to see where I reach practical limits of execution/liquidity with day trading. Like P2, I may have to move from 5m time frame to higher time frame if I ever go 'big time'. For now, though, it's a nice living. :)

dottom
 
Originally posted by Colonel_Blood
Another interesting point is this, the world record for the same colour coming up in a row on Roulette is about 48 in a Puerto Rican Casino in the 1950's, now what nobody can ever explain to me is that how come something like this (even taking the 0 or 00's out) can happen. Surely a run of at least 100 is possible?
More...
Crooked wheel?
 
Originally posted by darkhorse
The correct thing, in my opinion, is oddly as unpopular as the wrong technique is popular: to increase your position size into winning streaks. This may hurt you a little when those streaks are cut short- but the cost is light, just a shave off the top after already being ahead. And when you do catch an outlier streak in your favor, and have been gearing up to take full advantage of it- that's when you blow the doors off. And you can rest easy knowing it was an extra chunk of profit at risk, not initial capital.

Never let them get too deep into your pocket. Ever.
More...

Go with the trend... If your equity curve trend is down, back out of it.

cpo
 
You must log in or register to reply here.
Back
Top