Nice work. Care to link an xls version? What market(s) are the values comprised of?
You are using an out of date browser. It may not display this or other websites correctly.
You should upgrade or use an alternative browser.
You should upgrade or use an alternative browser.
Not Holy but Close Grail
- Thread starter crash n burn
- Start date
Quote from Michael Black:
You bring up an interesting concept. What other than the past do we have to make forecasts? I should think most actions we take by virtue of choice is dictated by the regularity of observed patterns.
So I think that so long as the frequency of regularity pasterns remain intact, and significant, what reason do we have to dismiss them?
My question is really an attempt to understand modeling time series. Although I believe the numbers do seem rather impressive, I know from experience that events can hog up a great portion of of the data and essentially cluster the impressive occurrences for that period.
I think of it like a coin in a coin toss game that at some point gets replaced by a biased coin for a portion of time. Seeing the final results many iterations later perhaps might show a distribution of heads that are at the 5 sigma level, but without seeing the tosses during the portion of time with the biased coin, one might presume a constant process, which of if the coin were switched back to the fair coin would be the wrong conclusion.
Michael, I have a few questions for you:
1) Could anyone predict the future?
2) Is it still safe to believe that the market will continue to go up like before?
3) If you have to make a forecast, wouldn't it be more relevant to use data closer to your period of forecast, which will take the current economic and market regime into consideration?
Thanks!
PA
Quote from Pension_Admin:
Michael, I have a few questions for you:
1) Could anyone predict the future?
2) Is it still safe to believe that the market will continue to go up like before?
3) If you have to make a forecast, wouldn't it be more relevant to use data closer to your period of forecast, which will take the current economic and market regime into consideration?
Thanks!
PA
1. Isn't that what we are trying to do? No forecast can be certain, but with a degree of confidence we can assume continuation of some pattern. e.g I expect to eat dinner tonight, but I could drop dead before the food has arrived.
2. No and no time in the past ( sections of 1 yr, 10 or whatever were the same.) With 100s of years of this capitalism system developing, basically a near random walk has a pos bias. What reason do I have to believe that process will change in the next 10 years?- The day results we are questioning, did they show an even distribution over the 100 yrs, say were the percentages similar over the decade of the 70s as say the 30s as say the 90s?
3. Yes; however it depends on the system one is modeling. If a pattern repeats, has a duration, the context, the order of it's expression may require a longer rather then a shorter time frame to realize.
So if I pose a question. Suppose the following hypothetical.
You have 2 trading systems, each containing 100,000 trades. Both systems have the exact same t score, lets say of 4. So T= avg trade/ std. dev of avg trade * N(100,000) trades. The difference is that system A performs this result over 1 year's time, System B 's results are for a 10 year period. Which system would you feel most comfortable trading next year?
Quote from Michael Black:
1. Isn't that what we are trying to do? No forecast can be certain, but with a degree of confidence we can assume continuation of some pattern. e.g I expect to eat dinner tonight, but I could drop dead before the food has arrived.
2. No and no time in the past ( sections of 1 yr, 10 or whatever were the same.) With 100s of years of this capitalism system developing, basically a near random walk has a pos bias. What reason do I have to believe that process will change in the next 10 years?- The day results we are questioning, did they show an even distribution over the 100 yrs, say were the percentages similar over the decade of the 70s as say the 30s as say the 90s?
3. Yes; however it depends on the system one is modeling. If a pattern repeats, has a duration, the context, the order of it's expression may require a longer rather then a shorter time frame to realize.
So if I pose a question. Suppose the following hypothetical.
You have 2 trading systems, each containing 100,000 trades. Both systems have the exact same t score, lets say of 4. So T= avg trade/ std. dev of avg trade * N(100,000) trades. The difference is that system A performs this result over 1 year's time, System B 's results are for a 10 year period. Which system would you feel most comfortable trading next year?
1. You are right. Predicting the future is what everyone is trying to do, but using the past to predict the future lack relevancy, especially when the statistical results are not explained.
However, once we figure out the reason for the results, predictive value may disappear--like TA.
2. The problem with this type of thinking is that it disregard how long an economic downturn could last, and the life expectancy of the investors/traders in relation to the market. The market may drift upward in the long-run, but what's matter to investors/traders is how the market move in their short lifespan.
Also, the upward drift in the market may not be caused by capitalism alone. There are many factors that cause an upward drift in the market, and these factors may or may not continue to be present as our political and economical climates change.
3. If the system is based on past 100 years of data, then I really couldn't see the correlation and the causation to the predicted outcome. I would be indifferent to using a coin toss or using something totally unrelated to the market to model the market.
To me, 50/50 is good enough.
To answer your question, I would be indifferent to both systems, because I don't see the correlation and the causation to the predicted outcomes in both systems.
PA
Sent you a PM. LMK if you have any questions.
Thank you for asking, but no. That would require a longer post than I would care to write. However, if you choose to test this nonsensical "almost grail" going forward, I think you will see for yourself. And should I be proven wrong after a fair number of trials, please be sure to point out how clueless I am. I think that the most basic elements of TA have practical value insofar as preceding price action is concerned. However, I'm inclined to believe that it is better to use surgical gloves rather than oversized clown mittens.Quote from Michael Black:
Please elaborate.