Day Trading Stocks & Futures

[ncube]

2008 like event just dn't happen out of the blue. Your system should see it coming.

I understand what you are saying, but my system is mostly long only & mechanical based only on price action. It does not use any fundamental/economic signals for decision making. Hence it is difficult to catch such situations. I tried few models to capture these economic & emotional trends but did not see much success.
We can also use some technical indicators say if the market goes below 50/100/200 EMA exit the market completely or go short, but this is not a robust design. Our system should be robust and optimal enough to handle all types of situations and give decent returns over the long run.

 

[manny.here]

Wishing everyone a very happy and profitable year 2020 to all fellow learners and experienced traders

 

[TracerBullet]

Thanks for the info, Actually Capital is not a constraint, on average I am only invest about 70% of my capital in equity rest 30% is in debt fund. Also we can easily achieve leverage target by using a combinations of Futures and options. However by choice I do not prefer leverage as my back testing shows that if there is 2008 like situation then with 3X leverage my portfolio will be down 90%... . Hence what I am planning to do is split a small percentage say 20% of my portfolio and leverage it 3X times which can easily double my total returns with nominal increase in risk.

I never really understood this thing - of marking a smaller part of capital with higher risk - maybe i missed something.
How is this any different than slightly increasing your risk per trade for all trades. And how will it improve your R:R ?

Assuming you get 20% returns with 10% dd again next year,
returns = 20*0.8 + 20*0.2*3 = 28
DD = 10*0.8 + 10*0.2*3 = 14

So how did it improve anything ?
-------------------
Eventually i would look at backtesting Daily/weekly too.
How did you manage your data? Did you get a clean db with dividends/splits/rightsissues/bonus/etc adjusted ? And also delisted stocks.
For American stocks, Norgate seems to have good reputation for getting this kind of data. Not sure about India. I would probably just clean up large changes and use bhavcopy but would be nice to hear how you test.

 

[lemondew]

bhai 28% is 40% more than 20 %. Also with 3x levrage in 2008 it would be down 90% with 1 x it will be 30% down and just 20% 3x his dd will be down. The same can be achieved by decreasing/increasing risk per trade.

Thats what i understand. Last part of your question regarding clean db would be good if someone has it.

I never really understood this thing - of marking a smaller part of capital with higher risk - maybe i missed something.
How is this any different than slightly increasing your risk per trade for all trades. And how will it improve your R:R ?

Assuming you get 20% returns with 10% dd again next year,
returns = 20*0.8 + 20*0.2*3 = 28
DD = 10*0.8 + 10*0.2*3 = 14

So how did it improve anything ?
-------------------
Eventually i would look at backtesting Daily/weekly too.
How did you manage your data? Did you get a clean db with dividends/splits/rightsissues/bonus/etc adjusted ? And also delisted stocks.
For American stocks, Norgate seems to have good reputation for getting this kind of data. Not sure about India. I would probably just clean up large changes and use bhavcopy but would be nice to hear how you test.

 

[siddhant4u]

people did see it coming and that is why they made huge money shorting the banks and housing market around the world.

Very few got into shorting subprime and minted money. By all accounts their numbers can be counted on fingers compare to millions of traders worldwide

 

[against_tides]

Happy New year Bhai log..cheers

 

[TracerBullet]

bhai 28% is 40% more than 20 %. Also with 3x levrage in 2008 it would be down 90% with 1 x it will be 30% down and just 20% 3x his dd will be down. The same can be achieved by decreasing/increasing risk per trade.

Yes, 14% is also 40% more. I can increase my risk and my dd as needed.

But there is no extra benefit of saying i take more risk in small capital and less risk in other for same system.
Its seems just a roundabout way of increasing your overall risk to 1.4x with proportionate increase in reward.
And it will probably create extra hassle in execution by dividing trades into two camps.
Anyway, this is how i think. Just asking what am i missing here ...

 

[AD_Trader]

Wishing fellow members a very happy new year. May this year bring lots of wealth, prosperity and happiness.

 

[ncube]

I never really understood this thing - of marking a smaller part of capital with higher risk - maybe i missed something.
How is this any different than slightly increasing your risk per trade for all trades. And how will it improve your R:R ?

Assuming you get 20% returns with 10% dd again next year,
returns = 20*0.8 + 20*0.2*3 = 28
DD = 10*0.8 + 10*0.2*3 = 14

So how did it improve anything ?
-------------------
Eventually i would look at backtesting Daily/weekly too.
How did you manage your data? Did you get a clean db with dividends/splits/rightsissues/bonus/etc adjusted ? And also delisted stocks.
For American stocks, Norgate seems to have good reputation for getting this kind of data. Not sure about India. I would probably just clean up large changes and use bhavcopy but would be nice to hear how you test.

Ok..basics first...Lets say I have 1L capital, now I want to leverage 20% of it to 3X...which means my total portfolio value should become 1.4 Lakh (80000 + 20000 * 3). Now Lets say I have 30 stocks in my portfolio...which means I need to leverage each stock 1.4 times. This is very difficult in India as the lot sizes are not flexible and its not possible to leverage it in smaller units.

The best alternative option then is to leverage the portfolio beta with index such as Nifty using Futures/Options. I.e I just need to measure my portfolio beta with Nifty and buy/sell Nifty futures/options in correct proportion.

By doing this if next year also I make 20% and 10% draw-down then my total returns would be as follows:

a. 20% of 80000 = 16000
b. 20% of 60000 = 12000
Total returns = 28000 = 28%

Max DD:
a. 10% of 80000 = 8000
b. 10% of 60000 = 6000
Total DD = 14000 = 14%

Advantages:
1. With the same capital of 1L, now my total returns is 28% instead of 20% and Max DD is 14% instead of 10%. I.e I have boosted up my returns by 8% by allowing the DD to increase by 4%. (Note for long only equity portfolios DD is mostly paper loss unless it is booked).
2. Testing this out in a smaller account would help me to control the leverage easily, and if things go wrong at max I will be losing 20% of my capital which is unlikely as probability of another 2008 type event is low..but I would not risk based on that knowledge...

Regarding the clean data for backtesting, when I had prepared by the database there were no good data vendors in India for clean data, so I had to curate the data myself...its not perfect I would mark it about 75-80% clean. Since my strategy does not have a strict requirement for clean data it is fine for me. But if you require clean data you may check the historical data provided by Zerodha, I think it should be reasonably clean.

Hope the leverage concept is clear... Happy New Year !!!

 

[TracerBullet]

Advantages:
1. With the same capital of 1L, now my total returns is 28% instead of 20% and Max DD is 14% instead of 10%. I.e I have boosted up my returns by 8% by allowing the DD to increase by 4%. (Note for long only equity portfolios DD is mostly paper loss unless it is booked).
2. Testing this out in a smaller account would help me to control the leverage easily, and if things go wrong at max I will be losing 20% of my capital which is unlikely as probability of another 2008 type event is low..but I would not risk based on that knowledge...

1) They are both 20% increases from base. Yes, returns is more than dd so you get 8% extra for 4% dd. So, The edge itself is same and you can get returns as per your risk tolerance. I understand taking extra risk for the short term to increase capital, but that does not increase Reward/Risk of your system. There is no doubling of reward without doubling of risk - for the same system.
We cannot become buy and hold from mechanical system, have to follow plan. The unknown risk ofc is that we can get worse dd than max in future / markets can change. Hence the need to keep it under control atleast in long term.

2) Yes but it will have same impact on your 80% account too - in proportion.

Anyway, the issue of min size is there, but you will have to then selectively increase size of some position. Otherwise leverage will be only on 20% which defeats the purpose.

Regarding the clean data for backtesting, when I had prepared by the database there were no good data vendors in India for clean data, so I had to curate the data myself...its not perfect I would mark it about 75-80% clean. Since my strategy does not have a strict requirement for clean data it is fine for me. But if you require clean data you may check the historical data provided by Zerodha, I think it should be reasonably clean.

Hope the leverage concept is clear... Happy New Year !!!

Yeah, that is what i will probably do too. Do you mean zerodha gives adjusted data for stocks, back adjusting corporate actions. There is also selection bias problem with longer holding systems, so we should have stocks that fade away too, Esp for long only systems.

I still dont see any special purpose of higher risk on small capital + normal risk on rest capital for same system - apart from possible execution issues .. Happy new year too all. Good luck