Discussion in 'Trading Diary' started by augubhai, Jun 25, 2013.

1. ### augubhaiWell-Known Member

Joined:
Jan 10, 2006
Messages:
8,662
8,631
Trophy Points:
113
Re: What I want to do...

Yes, I think so too. It has to be either scalpers, or quants running multiple algos across diversified markets.

So, I guess that I should be seriously looking at scalping, and try to find a system with negligible drawdown. Why be in love with drawdowns?

Tlahuicole and timepass like this.
2. ### augubhaiWell-Known Member

Joined:
Jan 10, 2006
Messages:
8,662
8,631
Trophy Points:
113
3. ### augubhaiWell-Known Member

Joined:
Jan 10, 2006
Messages:
8,662
8,631
Trophy Points:
113
Kelly Criterion on variable R:R systems

For 10 months now, I have been using the Kelly Criterion for position sizing. It's been working fine, and allows me to take bigger position sizes than I would have otherwise dared to. However, I always have my doubts on whether I am using it right.... especially when hit by severe bouts of volatility.

I have calculated the Kelly criterion based on backtests on historical data. Right now, I am even suspecting the validity of those backtests because of new factors 1. Impact of Banknifty weekly options 2. Shifting the RBI announcement time to 2:30 (Announcement days very profitable in the historical tests). For now, my trading continues to be based those backtests results.

Kelly % = W – [(1 – W) / R]
Where:
W = Winning probability
R = Win/loss ratio

W is ok. I just take the Win Rate after adjusting for costs and slippage.

R is a problem. The formula was designed for fixed risk-reward systems, whereas my systems have variable risk-reward. There may be studies done on applying this criterion for variable risk-reward that I am yet to know about.....

At the moment, these are the options that i can think of....
1. Average win/Average Loss (Payoff Ratio in Amibroker)
2. Average win/Average MAE
3. Average win/Average Initial Risk

Option 1 is what I have been using until now. Going forward, to reduce system volatility, while at the same time not discarding potentially good systems, I am thinking of using the average of options 1 and 2.

What am I doing right or wrong?

Last edited: Jun 28, 2017
4. ### tkamalMember

Joined:
May 22, 2007
Messages:
78
58
Trophy Points:
18
Respected Augubhai,
We all know you are successful and experienced intraday trader. I need your guidance regarding this.>> For Intraday trade (in equities or futures ) (stocks only) what is the more safer stoploss and target for maximum win strike?
Due to job time restriction I can trade only during 9.00 AM to 10.30 AM only.
Thanks sir.
-tkamal

augubhai likes this.
5. ### augubhaiWell-Known Member

Joined:
Jan 10, 2006
Messages:
8,662
8,631
Trophy Points:
113
Hi tkamal,

If I knew the exact answer to your question, then I would be a very rich man, but I am not a very rich man.

When i first started intraday trading, I too had the same constraints as you. In those days, NSE used to open at 9:55 AM. So I traded the Opening Range Breakout method by placing Buy and Sell stops at the High and Low of the first 5 minute bar, and then would rush to office. Later, I would check the trade occasionally from my office to modify Stop Losses, if needed. I would close the trade before 3:30 PM. I mostly traded this method on Nifty futures.

I have shared the system here: http://www.traderji.com/amibroker/52959-augubhais-orb-system.html

Rules:
1. Place Buy Stop at the first 5 min bar high, and Sell Stop at the first 5 min bar low.
2. If either order is executed, modify the other order and set Stop Loss at 0.614% from the entry trigger price.
3. If the Stop Loss is not hit, square of the position at the end of day.

Surprisingly, this method worked very well for a long time, but in the end it did not make me a rich man yet. So what is the use?

vijkris and tkamal like this.
6. ### copypasteaeeWell-Known Member

Joined:
Aug 15, 2014
Messages:
748
881
Trophy Points:
93
Re: Kelly Criterion on variable R:R systems

You can also use profit factor. Total profit/ total loss

augubhai likes this.

7. ### copypasteaeeWell-Known Member

Joined:
Aug 15, 2014
Messages:
748
881
Trophy Points:
93
Howdo you calculate lot size using Kelly criterion.

augubhai likes this.
8. ### augubhaiWell-Known Member

Joined:
Jan 10, 2006
Messages:
8,662
8,631
Trophy Points:
113
Profit Factor is not relevant for Kelly Criterion calculation. Kelly Criterion is used to find the optimal percentage of capital to be risked to maximize long term gain. It depends on only 2 factors - Win Rate and Risk-Reward ratio.

9. ### tkamalMember

Joined:
May 22, 2007
Messages:
78
58
Trophy Points:
18
Many Thank you sir for your useful guidance..
Good day sir...

-tkamal

Joined:
Jan 10, 2006
Messages:
8,662