LOL Not for anything Abhay, but if you already KNOW the market in advance then why are you bothering to backtest a simple system such as this??? If you already KNOW then why rely on probabilities and odds? PEACE and good-trading, Commisso EDIT; I already feel dumber for having replied to you...
Actually, this is a valid question. If I were thinking about utilizing your service, this thread would give me pause.
i spent some time tracking mohan for a friend. save your money, he's not worth it, the daily updates are not all even written in real time. there are better traders here and on #indexfutures.
Seriously Abhay if you are trying to use ET to push your 'Miss Cleo' service to scarecrows then perhaps you should make a new handle (name it applehead) for your threads such as this...
Abhay, I have done work based on your original inquiry. I have found that the probability distribution for your inquiry is closely related to the probability distribution of a random-walk process for n steps. For example, I studied the following question: If a given security had a closing price of 10.00 on 'day 1', and a closing price of 10.30 'day 2', what is the probability that 'day 6' would have a higher closing price than 'day 1'? Relative to day 1, I counted (well, I didn't count it myself, software did) the number downticks over the next 4 days. Ultimatley, it created a nice combinatorial model which captured all possible outcomes (0 downticks, 1 downtick, 2 downticks, 3 downticks, 4 downticks) over a 4 day period. The results I came up with were very close to theoretical results (i.e. if you were to just use probability theory to solve this problem). If this doesn't make any sense, essentially I showed that if a stock price went up every day for four (4) days, there is no statistical advantage to purchasing that stock - it's price might be higer in 5 days in the future, it might be lower in 5 days in the future. I studied 500 different companies, with all available historical data for each company (or, about 4,000 years of data in total). Regards, Brandon
Let me say, I believe this will work best on index Specially these parameters are suited to S&P index We can define similar parameters for DOW too. For those asking why I am pondering about this, because I always continue thinking. If any idea comes to my mind I like to research it out. Even my forecasts can help in identifying the direction bias I should be taking, or if I see a choppy day I may better change my strategy for that day. Direction bias from forecasts is a big ++ Rest all I was doing was asking for some help in research and Electric Savant very nicely defined it further....
Well, FABULOUS. Then you go first, and tell us how you've done in a couple months. We'll be here when you get back.
Before jumping in this I need to statistically test this and tweak the parameters... *lol* I am trying to write the strategy myself too and any help or verification by others will certainly help. I have esignal and it has only 3-4 months of intra-day data
Abhay, I have found that you only need about 2 years of ES backtesting. You might consider SP backtesting to avoid the spikes that the ES gives. You also might discover the last 10 months to be more successful with your idea. You must eventually get an adaptive addition to your statistics utilizing a measuremet of the range of the market. Michael B.