I guess I'll read more
No one is talking to me. No one is even downloading my files...
Sucks.
I got yelled at today at work. When I told him why was he yelling, it turned out he wasn't mad at me, and he apologized. But obviously he was indeed yelling at me. ****er.
It's because I helped someone, then that someone said that the others weren't doing their job (because I had helped him so much), and the boss of that group wanted to take it out on me. I said "fine, from now on, anyone asking me for help, I'll send them straight to you, and refuse to help at all". Like everyone else does at my office. Send the problem to someone else. As we say in Italy, "chi non fa non sbaglia". Those who don't do cannot make any mistakes. Which means that making mistakes is a sign that you are doing things.
I am going to read some more book, or else I'll get lazy and stop reading it. I know how things go.
page 81, OTHER STATISTICAL TECHNIQUES AND THEIR USE
Another day of quiet Vito at work. As long as I don't lower my guard, things should be easier from here on. I don't talk to him, I treat him coldly, I help him if he needs help, and I get respected. If I treat him nicely instead he starts busting my balls very quickly.
The following section is intended only to acquaint the reader with some other statistical
techniques that are available. We strongly suggest that a more thorough study
be undertaken by those serious about developing and evaluating trading systems.
Oh ok, they acquaint me but I have to do more work on it, on my own. 400 pages is not enough.
Page 81, Genetically Evolved Systems
We develop many systems using genetic algorithms. A popular fitness function (criterion
used to determine whether a model is producing the desired outcome) is the
total net profit of the system. However, net profit is not the best measure of system
quality!
Yeah, this is great. Because thanks to RiskOptimizer now I am totally familiar with genetic algorithms. They go where a brute-force optimizer cannot go, because it cannot try trillions of solutions.
But RiskOptimizer is a tool for portfolio optimization. I still do not know if and how I could use it for system creation. However, I do have an idea. If I loaded the data for a system on excel, and created entries and exits functions, I could easily have RiskOptimizer adjust those cells. The problem is that it is not that easy to back-test systems on excel. But soon I'll be familiar with that program enough to do that, too. I think RiskOptimizer can really take me far. I could even write a function on vito and have RiskOptimizer adjust my relationship with him. The constraints would be that I have to stay in the same room with him and that I cannot kill him or attack him physically. The other cells could all be adjusted. The cell to be optimized would be disturbance: "reduce the amount of disturbance to the lowest", given these constraints and with disturbance being a function of so and so. I suppose I've already found - but after 3 months - the optimized values. They have to do with our conversations, and precisely I have to answer "yes" and "no" to all his questions, with a lower and lower volume, until I almost don't answer anymore. This brings disturbance from vito to optimal (low) levels.
A system that only trades the major crashes on the S&P 500 will yield a
very high total net profit with a very high percentage of winning trades. But who
knows if such a system would hold up? Intuitively, if the system only took two or
three trades in 10 years, the probability seems very low that it would continue to
perform well in the future or even take any more trades. Part of the problem is that
net profit does not consider the number of trades taken or their variability.
Yeah, who knows if this system will hold up? I haven't backtested it. But I've talked to people about it, and they agree with it. I need the carrot, too, though. I do use it already. If he asks for help, I promptly help him. To show him that I am fair, and that if he's serious, I am willing to interact with him.
An alternative fitness function that avoids some of the problems associated
with net profit is the t-statistic or its associated probability. When using the t-statistic
as a fitness function, instead of merely trying to evolve the most profitable
systems, the intention is to genetically evolve systems that have the greatest likelihood
of being profitable in the future or, equivalently, that have the least likelihood
of being profitable merely due to chance or curve-fitting. This approach
works fairly well.
Yeah, I wish I could do this stuff. There's too many formulas for me to figure that much out. On the other hand, if I quit my job, a lot of intellectual resources would be freed up.
The t-statistic factors in profitability, sample size, and number
of trades taken. All things being equal, the greater the number of trades a system
takes, the greater the t-statistic and the more likely it will hold up in the future.
Oh, I know this much. I'm already using this knowledge in building my systems. It is self-intuitive.
Likewise, systems that produce more consistently profitable trades with less variation
are more desirable than systems that produce wildly varying trades and will
yield higher t-statistic values. The t-statistic incorporates many of the features that
define the quality of a trading model into one number that can be maximized by a
genetic algorithm.
Yeah, that's nice. Still beyond my reach though.
Page 82, Multiple Regression
http://dictionary.reference.com/browse/regression
statistics
a. the analysis or measure of the association between one variable (the dependent variable) and one or more other variables (the independent variables), usually formulated in an equation in which the independent variables have parametric coefficients, which may enable future values of the dependent variable to be predicted
Not enough. This is still Greek to me.
Another statistical technique frequently used is multiple regression. Consider
intermarket analysis: The purpose of intermarket analysis is to find measures of
behaviors in other markets that are predictive of the future behavior of the market
being studied. Running various regressions is an appropriate technique for analyzing
such potential relationships; moreover, there are excellent statistics to use
for testing and setting confidence intervals on the correlations and regression
(beta) weights generated by the analyses. Due to lack of space and the limited
scope of this chapter, no examples are presented, but the reader is referred to
Myers (1986), a good basic text on multiple regression.
Oh yeah, sure - i just have to read another book.
Monte Carlo Simulations
One powerful, unique approach to making statistical inferences is known as the
Monte Carlo Simulation, which involves repeated tests on synthetic data that are
constructed to have the properties of samples taken from a random population.
Except for randomness, the synthetic data are constructed to have the basic characteristics
of the population from which the real sample was drawn and about
which inferences must be made. This is a very powerful method. The beauty of
Monte Carlo Simulations is that they can be performed in a way that avoids the
dangers of assumptions (such as that of the normal distribution) being violated,
which would lead to untrustworthy results.
Yeah, this is what I am doing. That software had "monte carlo" written all over it. Yeah, from the way they talk about it, it sounds like it would have the advantages of using an out-of-sample that takes little bits of data from the whole sample rather than from the end of it.
Out-of-Sample Testing
Another way to evaluate a system is to perform out-of-sample testing. Several time
periods are reserved to test a model that has been developed or optimized on some
other time period. Out-of-sample testing helps determine how the model behaves
on data it had not seen during optimization or development. This approach is
strongly recommended. In fact, in the examples discussed above, both in-sample
and out-of-sample tests were analyzed. No corrections to the statistics for the
process of optimization are necessary in out-of-sample testing. Out-of-sample and
multiple-sample tests may also provide some information on whether the market
has changed its behavior over various periods of time.
The good thing about the investors is that they sugar-coated the out-of-sample for me, because had I learned it from a book, I never would have listened to the book about it, as it is also telling two hundred other different things and it says they all might be good. Damn. These comprehensive books suck. The investors just told me: why don't you use out-of-sample? And since then I've been using it. This book wrote dozens of pages about out-of-sample testing but also about all the other types of methods, and of course your reaction is to reject the whole lot of them. That's why you learn more from one sentence on a forum than from a whole textbook.
On the other hand, I owe RiskOptimizer to this book.
page 83, Walk-Forward Testing
Here they go again, bringing it up one more time. They should have focused on this all at once. Anyway, let's hear the same things all over again:
In walk-forward testing, a system is optimized on several years of data and then
traded the next year. The system is then reoptimized on several more years of data,
moving the window forward to include the year just traded. The system is then
traded for another year. This process is repeated again and again, “walking forward”
through the data series.
Yeah, I knew this.
Although very computationally intensive, this is an
excellent way to study and test a trading system. In a sense, even though optimization
is occurring, all trades are taken on what is essentially out-of-sample test
data. All the statistics discussed above, such as the t-tests, can be used on walkforward
test results in a simple manner that does not require any corrections for
optimization. In addition, the tests will very closely simulate the process that
occurs during real trading--first optimization occurs, next the system is traded on
data not used during the optimization, and then every so often the system is reoptimized
to update it. Sophisticated developers can build the optimization process
into the system, producing what might be called an “adaptive” trading model.
Meyers (1997) wrote an article illustrating the process of walk-forward testing.
Hmm, I don't know about this. More complex stuff thrown at me, to screw up my work. Let's keep things simple. Forget this walk-forward crap. If my systems become obsolete, I'll just keep on building new ones, at the rate of 20 per year. And I'll keep on replacing those that have stopped being profitable (i.e.: exceeded their max drawdown or similar).
CONCLUSION
In the course of developing trading systems, statistics help the trader quickly reject
models exhibiting behavior that could have been due to chance or to excessive
curve-fitting on an inadequately sized sample. Probabilities can be estimated, and
if it is found that there is only a very small probability that a model’s performance
could be due to chance alone, then the trader can feel more confident when actually
trading the model.
Yeah, as I said, I have no idea how to measure this correctly. I'll use my usual guesstimates.
There are many ways for the trader to use and calculate statistics. The central
theme is the attempt to make inferences about a population on the basis of
samples drawn from that population.
Keep in mind that when using statistics on the kinds of data faced by traders,
certain assumptions will be violated. For practical purposes, some of the violations
may not be too critical; thanks to the Central Limit Theorem, data that are not normally
distributed can usually be analyzed adequately for most needs. Other violations
that are more serious (e.g., ones involving serial dependence) do need to be
taken into account, but rough-and-ready rules may be used to reckon corrections
to the probabilities.
Yeah, I have no idea what
violations they are talking about.
The bottom line: It is better to operate with some information,
even knowing that some assumptions may be violated, than to operate blindly.
We have glossed over many of the details, definitions, and reasons behind the
statistics discussed above. Again, the intention was merely to acquaint the reader
with some of the more frequently used applications. We suggest that any committed
trader obtain and study some good basic texts on statistical techniques.
Oh yeah? How much more do I have to read then? Let's keep going. Stop telling me to get more books and read them.
Page 85, The Study of Entries
This is a whole new section, Part 3. I'll have to read the whole section, because they're not talking about entries, except in the first three pages. They're basically talking about different strategies, which is what I started reading this book for, which, after all, is entitled: "The Encyclopedia of Trading Strategies", and that is why I came across it on the web.
This is finally it, after 85 pages. I will resume from here tomorrow. Knowing myself, while reading the next chapter, I'll be stopping and building a few dozen new systems along the way. I've built most of my systems without any inputs from the outside world. Imagine what happens if I actually start listening to others.