Prudent Risk Management + No Edge = Positive Expectancy??

[Zwaen]

Why do you need Kalman Filter here? You are only using GBM to generate a sim stock price and not doing any prediction or forecasting?

I think one can use a Kalman filter to stay in the 'trend', and thus has the same function as a trailing stop. )experimented some with it, but isn't any more workable then a simple ma)

 

[Zwaen]

Exactly, there is no edge in trend trading. You can close your eyes, flip a coin, scratch your left nut to decide when to buy or sell, and cut short you losers and let your profits run. In a trend, any of those entries would work.

I think retail can utilise the only edge they have above institutions, getting in and out in a few seconds. This is one of the few (if any) edges which is 'real' imo.

 

[ironchef]

I think retail can utilise the only edge they have above institutions, getting in and out in a few seconds. This is one of the few (if any) edges which is 'real' imo.

I think our real advantage is we don't report to anyone, no OPM to worry, customers pestering us and we can stay out if we are not happy with the market... We can focus on long term, no one fires us if we have short term losses....

A pro cannot not trade or stay out for too long, they need to show short term results or be looking for another job.

 

[ironchef]

https://www.elitetrader.com/et/threads/flip-of-a-coin-but.372081/

Thank you. You are the math wizard I have been looking for.

Where have you been? I have been looking for this for quite some time, how could I have missed this?

Then I looked at the date. I quitted ET and most trading for two years to take care of a family health issue. Came back to ET in early 2023 but this was posted before my return.

 

[smacdtrader]

I have been trading a system for 7 weeks with positive results.

The system is eerily similar to a random coin toss, win rate ~ 50%, R:R ~ 1:1 but both are skewed slightly favoring win and reward. Statistical sample analysis showed these are not statistically significant, i.e., could be due to sample size errors.

Risk Management: Cut losses aggressively, take wins aggressively. Exit trades, call it a day when cum trades produce a profit or total # of trades exceed X, which ever comes first.

75% winning days and 6 consecutive winning weeks are highly unusual for a random coin toss.

Could @Buy1Sell2 be right? Or I am fooled by randomness?

Comments/feedbacks are welcome especially from the math wizards on ET.

How many trades total is more relevant than how many weeks? If you only have like 10 or 20 trades, then this is well within the boundaries of a random walk, which is what you're describing I think. But if you've got closer to 50 or 100 trades then you may have something. Statistically speaking if it is a coin toss to win or lose, and the w/l ratio is even slightly greater than 1, then given enough opportunities you will make a bag. Mathematically the limit goes to infinity.

 

[smacdtrader]

No commissions trading stocks.

As of last week, after 397 trades, 215 win. R:R more difficult to determine, perhaps 1.05:1

Last week is more difficult to assess. I am experimenting with not exiting after cum win but trade a fixed number everyday. Preliminary outcome: Not good. Two losing days in a row were losses when I made 50 trades each day.

Not scalping, trend following.

dude, this is way better than a coin toss and 1:1. be precise man. if you can do that trade as much as possible its worth millions and millions.

 

[smacdtrader]

What is risk management but a prediction of the future? If you can’t predict the future with any statistical significance you can’t do “prudent risk management”

imo, risk management is accepting you haven't and can't predict the future and preparing for what you aren't sure will happen.

 

[smacdtrader]

I don’t believe stock price movement is random.

By definition of random, it is random, at least from a probability standpoint. Stock price is a pre-image mapping of a measurable space in a sample space, random. Colloquially, its also pretty uncertain to most of us, so even if it's totally manipulated by someone, our experience is that it's random af.

 

[ironchef]

How many trades total is more relevant than how many weeks? If you only have like 10 or 20 trades, then this is well within the boundaries of a random walk, which is what you're describing I think. But if you've got closer to 50 or 100 trades then you may have something. Statistically speaking if it is a coin toss to win or lose, and the w/l ratio is even slightly greater than 1, then given enough opportunities you will make a bag. Mathematically the limit goes to infinity.

I understand that. The $64,000 question is the win/lose ratio greater than 1, do I have enough sample size to definitively say it is not random.

 

[ironchef]

dude, this is way better than a coin toss and 1:1. be precise man. if you can do that trade as much as possible its worth millions and millions.

I know you are skeptical, I am too. Most likely it is an illusion.

When it is too good to be true, it is not true.

By definition of random, it is random, at least from a probability standpoint. Stock price is a pre-image mapping of a measurable space in a sample space, random. Colloquially, its also pretty uncertain to most of us, so even if it's totally manipulated by someone, our experience is that it's random af.

I think you understand statistics and probability better than most of us here.