Thanks for the response. I'd be interested in seeing your thought process as you work through this issue.
Taking 320K to 3.5million by Year End 2009
- Thread starter neke
- Start date
- Status
Thanks for the response. I'd be interested in seeing your thought process as you work through this issue.
After some thought, the fraction, f, to use on a trade shall be a function of 3 variables for stocks, and 4 variables for options:
f = g(cratio, histf(strategy), recentf(strategy)) --stock
f = g(cratio, histf(strategy), recentf(strategy), daystoexp) --options
The meaning of each variable is below:
cratio
This is the ratio of current account value at the time of the trade to the starting value of the thread (320K). The idea is to increase fraction used as the buffer increases, subject to not exceeding the optimal. So at the start of the thread, cratio was 1 for the first trade. At the moment it is about 0.94. The methods calls for extra-conservatism when the fraction is low.
histf(strategy)
This is the historical optimizing fraction for the strategy, which is recalculated after each trade using the strategy. (Example if a strategy has two trades: bought at 10, sold at 12 in the first trade, and then bought at 120, sold at 102 for the second trade, that is 20% gain followed by 15% loss in the underlying, the optimal fraction - what maximizes the total return compounded - is 0.83. This means total return is maximized by using 83% of account value per trade). There is an absolute maximum of 3.0 for stocks and 0.80 for options (arbitrarily chosen). If optimal fraction is negative, a value of 0 will be used. Adjustments will be made where there are not enough historical data (10 is my minimum) for a credible fraction. Fortunately almost all my strategies have enough historical trades to pass that test. The higher this value, the higher the value of f and vice versa.
recentf(strategy)
This is the recent optimizing fraction for the strategy, also recalculated after each trade based on the strategy. This only takes into account recent trades (I am leaning on using only the three most recent trades). If the most recent trade reveal a 0 optimal (negative or zero edge), the value of f will go down. Conversely the higher the value. The absolute maximum imposed by the prior paragraph applies here too.
daystoexp
For options the days to expiration determines what fraction that can be used. The closer to expiration of the option, the faster the percentage moves on the option premium, consequently the smaller shall be the value of f.
Next I shall look at my results for last year,and try to figure out a good formula for the function g as a combination of the above variables.
f = g(cratio, histf(strategy), recentf(strategy)) --stock
f = g(cratio, histf(strategy), recentf(strategy), daystoexp) --options
The meaning of each variable is below:
cratio
This is the ratio of current account value at the time of the trade to the starting value of the thread (320K). The idea is to increase fraction used as the buffer increases, subject to not exceeding the optimal. So at the start of the thread, cratio was 1 for the first trade. At the moment it is about 0.94. The methods calls for extra-conservatism when the fraction is low.
histf(strategy)
This is the historical optimizing fraction for the strategy, which is recalculated after each trade using the strategy. (Example if a strategy has two trades: bought at 10, sold at 12 in the first trade, and then bought at 120, sold at 102 for the second trade, that is 20% gain followed by 15% loss in the underlying, the optimal fraction - what maximizes the total return compounded - is 0.83. This means total return is maximized by using 83% of account value per trade). There is an absolute maximum of 3.0 for stocks and 0.80 for options (arbitrarily chosen). If optimal fraction is negative, a value of 0 will be used. Adjustments will be made where there are not enough historical data (10 is my minimum) for a credible fraction. Fortunately almost all my strategies have enough historical trades to pass that test. The higher this value, the higher the value of f and vice versa.
recentf(strategy)
This is the recent optimizing fraction for the strategy, also recalculated after each trade based on the strategy. This only takes into account recent trades (I am leaning on using only the three most recent trades). If the most recent trade reveal a 0 optimal (negative or zero edge), the value of f will go down. Conversely the higher the value. The absolute maximum imposed by the prior paragraph applies here too.
daystoexp
For options the days to expiration determines what fraction that can be used. The closer to expiration of the option, the faster the percentage moves on the option premium, consequently the smaller shall be the value of f.
Next I shall look at my results for last year,and try to figure out a good formula for the function g as a combination of the above variables.
Its time to get out now. Your down 18 grand and next week might be another 18 grand. Next thing you know you will be like Lenny Dykstra getting sued by the brokerages because you blew the account and couldnt repay the margin loan.
Quote from neke:
Weekly Update for week 1/50 ended 1/17/2009
Ugly week, down 19K (6%). Not exactly how I wanted to start the year.
The major setback was on Monday when I tried to fade the sell off. Too leveraged: kind of forgot I have no buffer, as I was starting with a fresh capital base. Tried to come back later in the week on reduced leverage. Really putting in a lot of work into deriving a formula that should govern my leverage. It should depend on strategy, historical performance of strategy, buffer above starting capital, and recent performance of strategy.
Code:Opening Balance: 320,064 Net gain for the week -18,534 ------------------------------------------------ Net Balance: 301,530 Number of Trades 23 Number of Profitable Trades 11 Top/Bottom Discretionary Trades for the week TICKER ENTRY DATE/TIME EXIT DATE/TIME QTY PURCHASE AMT SOLD AMT GAIN/LOSS TYPE POT 2009-01-15-10-33-35 2009-01-15-14-01-38 3000 195480 206279 10784 LONG SZCAD 2009-01-15-10-36-40 2009-01-15-14-02-34 10000 15400 22500 6936 SPY CALL ----------------------------------------------------------------------- SPY 2009-01-12-09-45-51 2009-01-12-15-43-09 7000 616843 606830 -10030 LONG SZCAJ 2009-01-12-10-08-27 2009-01-12-15-42-35 35000 58138 38500 -20191 SPY CALL
Why do you insist on trolling every single thread on ET? You're pathetic.
Hi neke,
Just a reminder. You owe IRS $4063.58 transaction tax for the week ended 1/17/2009 based on your trading amount (not including your option trading) of $1,625,432.00.
Hope this will never happen to you and to every traders. But it looks like the current situation is in favor of this new tax. It will be interesting to see how you can survive while this new law is passed. I will stop trading immediately and close my brokerage account when this happens.
Gook luck
Just a reminder. You owe IRS $4063.58 transaction tax for the week ended 1/17/2009 based on your trading amount (not including your option trading) of $1,625,432.00.
Hope this will never happen to you and to every traders. But it looks like the current situation is in favor of this new tax. It will be interesting to see how you can survive while this new law is passed. I will stop trading immediately and close my brokerage account when this happens.
Gook luck
While most are dismissing the above, it really should be cause for concern for everyone, even those who think it won't apply to them.
Put the idiot on ignore. Stop quoting him - so that *I* won't have to be subjected to his drivel.
Actually that is for only 4 trades out of 23. The total figure would be much higher. Let's just hope they let the sleeping dog lie. The loss of jobs/income tax from traders and brokerage houses should be much (Last year my commissions was $64K, enough to pay an Ameritrade staffer). If anyone should think of any rational reason for such a tax, it should be imposed when the economy is hot, not when it is dithering on the edge!
- Status