|
TranceAddict Investors Club @ Marketocracy (pg. 127)
|
View this Thread in Original format
| Shakka |
| quote: | Originally posted by Capitalizt
and unpredictable human behavior. No formula can accurately predict this stuff. |
Not to mention often irrational behavior. I think one can make generally close predictions about certain things, however I agree with Cap that a mathematical formula is not going to be particularly helpful for macro predictions that can translate into stock picking acumen. And one must NEVER forget that the stock market and the economy are two completely different animals that often swim in completely opposite directions. |
|
|
| Krypton |
| quote: | Originally posted by Capitalizt
Your forumula making skills will eventually make you rich in another area krypt, but I personally think forumlas are pointless when determining stuff like this. The market is influenced by constantly changing fundamentals, outside forces, and unpredictable human behavior. No formula can accurately predict this stuff.
If you can create a forumula first, THEN backtest it and find it to be historically accurate, I might take a second look. It's probably impossible though. Markets aren't 100% logical. |
Well, the purpose of the formula is to indicate the equilibrium price of the index. The expected value of the index. I'm not one to deny that market prices are constantly fluctuating, but I do stress a difference between market prices and market values. Market prices which are substantially disconnected from true market values indicates a major price movement is to be expected, because it is my belief that market value is inherent in the market price, no matter how disconnected they are. This is very useful because it enables one to expect a major market shift before it actually happens.
Now, as for backtesting. Check this out. My formulas are all within 3% of the actual market price. I consider this a very positive and successful test. It indicates to me the markets have all reverted back to their equilibrium price level, or their true value.
EDIT: Test is actually a failure.
 |
|
|
| Capitalizt |
So what exactly went into the forumulas krypt? Backtesting will obviously work if you based them on historical price movement ;)
I still don't think any forumula can predict where the market is going. What if mark to market accounting is eliminated Monday morning? What if Obama raises the capital gains tax rate to 40%? What if Israel goes to war with Iran? There's just too much non-math that effects markets for any forumla to be accurate. |
|
|
| jerZ07002 |
| quote: | Originally posted by Krypton
So according to my calculations, the 3 major indexes of the US market are at equilibrium, and thus, technically not cheap. This really irks me because I believe the markets are extremely cheap. The ratio between dividend/earnings yields to the 10-Year Treasury Yield is hugely in favor of a cheap market and expensive bond.
Any thoughts? |
apparently you don't trust your own calculations! |
|
|
| jerZ07002 |
| quote: | Originally posted by Capitalizt
Your forumula making skills will eventually make you rich in another area krypt, but I personally think forumlas are pointless when determining stuff like this. |
you should see the formulas I use to make tax calculations for future periods (mostly expense allocations and foreign tax credit calculations), and the amounts my firm charges for them (hint: > $400 an hour in certain cases). |
|
|
| Krypton |
| quote: | Originally posted by jerZ07002
apparently you don't trust your own calculations! |
No, I don't. I'v been tweaking them for the past 3 days. I'm trying to combine linear and exponential functions into one formula. It's hard work. |
|
|
| Krypton |
| quote: | Originally posted by Capitalizt
So what exactly went into the forumulas krypt? Backtesting will obviously work if you based them on historical price movement ;)
I still don't think any forumula can predict where the market is going. What if mark to market accounting is eliminated Monday morning? What if Obama raises the capital gains tax rate to 40%? What if Israel goes to war with Iran? There's just too much non-math that effects markets for any forumla to be accurate. |
The thing is, I can't rely solely on a linear or exponential function. Because the market isn't solely linear, or solely, exponential. It seems like a combo of both. As for the non-mathematical factors of the market, this formula is not attempting to estimate any of that. It's simply to establish an equilibrium price level for this year, and estimate into the future. So that, if the market price is above the equilibrium price, it indicates an expensive market, and therefore, I should limit my stock exposure. Below equilibrium, the market is cheap. I never try to estimate what the market will be x years from now. |
|
|
| Krypton |
| quote: | Originally posted by jerZ07002
you should see the formulas I use to make tax calculations for future periods (mostly expense allocations and foreign tax credit calculations), and the amounts my firm charges for them (hint: > $400 an hour in certain cases). |
are they slope formulas or exponential formulas? |
|
|
| Krypton |
Here is a chart comparing the exponential function to the linear function. Just for clarification, the formulas used are below...
Exponential function = y(t) = 8001(1.06^t)
Linear function = y(t) = 96.04t + 8001
Here is a graph based on the above stat chart. Which line looks the best? I'm probably going to create some formula in which a range is given, instead of an exact estimate. The exponential function predicts the DJIA at equilibrium 10,000 in 2013. The linear function predicts 10,000 at 2030, which is way too long. The average between the two says 10,000 should be reached in 2015. Any thoughts?
 |
|
|
| Capitalizt |
wow so according to the linear equation we're barely going to crack 10k by 2030? I'm a big bear, but even I'm not that pessimistic. :)
My gut feeling says your average (green line) will be accurate within +/-10%. |
|
|
| Krypton |
| quote: | Originally posted by Capitalizt
wow so according to the linear equation we're barely going to crack 10k by 2030? I'm a big bear, but even I'm not that pessimistic. :)
My gut feeling says your average (green line) will be accurate within +/-10%. |
I'm thinking about making my expectation window the space between the blue and green line. |
|
|
| jerZ07002 |
| quote: | Originally posted by Krypton
are they slope formulas or exponential formulas? |
they aren't strictly mathematical formulae. The problem I usually face with excel is how to create an algebraic formula while at the same time searching multiple data sets for that information I want to use in the formula. In any given formula I could have multiple embedded functions searching for information based on certain defined variables, all while conducting a mathematical equation. In my field, the mathematics are much easier than sorting out the information needed in the equations.
Another significant issue in tax calculations is developing functions that look to the past while doing a current year calculation. While that may sound easy in theory, in practice, it can be quite difficult to perform in excel. A significant complication is also ordering rules in tax law that allocate losses and credits in specific ways (i.e., categories of income, etc...). It's tricky ! So, its not about forcasting through developing new formulae. It's more about manipulating excel functions to create the math equation that the IRS regulations specify. |
|
|
|
|
