Search results

Hi, guys I asked this question in Software but I think that might be the wrong forum; no takers. Hopefully this is the right forum. I'm trying to calculate an equity column for an automated trading system in a spreadsheet. Column G is the adjusted close (yahoo eod data), column K is my...

Don't hold back, tell us how you really feel! :D :D TA purely and simply is just analysis of price action and volume action. A great deal of that "analysis" is just as you described. But not everybody looking at price action and volume action is an innumerate idjit. :cool:

Post-optimization is an attempt to solve a dynamic problem with a static solution. That's called "inherently sub-optimal". :D :D

Why would discretionary traders need math skills? I thought you guys did it by magic or psychic powers or something. :D :D Mechanical traders, on the other hand ... :cool:

According to a lot of the reviews at amazon, this book is crap. If there are blank charts that should have data and somebody actually advocating displaced(!) MAs of all things, I have to agree.

http://www.jurikres.com/catalog/ms_ama.htm#top

It's Jurik, not Juric, and I'm pretty sure it's not FIR. It seems to be like John's filter: first-order IIR, and if I had to guess, I would guess dual-adaptive. That's just a WAG, don't ask me to justify it. :D I will say that Jurik Research seems to produce the best blackbox technical...

http://mathworld.wolfram.com/MovingMedian.html

Using a simple moving average to remove the noise from your data is like using a meat grinder to remove the fat from your steak. :D Median filters have less variance for the same amount of lag. :cool:

It's real simple (in principle). To have a worthwhile trading system, it has to do two things. First, it has to make a profit AFTER fees, slippage, etc. Second, it has to beat buy-and-hold, otherwise why bother? What you need is a sufficiently long backtest that produces the number of...

I would hesitate to trade a market-specific strategy unless I backtested it completely and found it superior to the best general system I had available. Let me emphasize that I'm not saying robustness is preferred to actual winning results. Nothing succeeds like success. :)

You can backtest a strategy on any data as long as you're not just seeking to curve-fit the data. The goal is to have a robust system: one that isn't just some ad hoc exploitation of a particular set of data. If your system can't trade in more than one timeframe, that's a good indication that...

Not necessarily. The point of backtesting is to test your system, not to test your data. Even if the historical data is flawed, you can still get a good idea of your system performance if you do a proper backtest. Of course you'll have to accept the fact of that particular data source being...

Yes. It isn't time, it's trades. The number of trades in your backtest minus the number of indicator parameters you're optimizing has to be 30 or greater for you to be able to say anything statistically valid about your system performance. That's all. :cool:

Filtering isn't stupid at all. My TA totally depends on it. The question is, is a given filter based on reasonable or unreasonable assumptions? You corrected me when I said the Kalman filter used in the first article was linear. That's cool (and thanks :)), but the ukf still depends on a...

I was fascinated by the fact that although math whizs Eckhardt and Trout disagreed on the value of moving averages (Trout was right :D), they both agreed that RSI, stochastics, Gann angles and Fibonacci anything was all garbage. Always nice to have my gut feelings confirmed by the experts. :D :D

"Everything should be made as simple as possible, but not simpler." ~ Albert Einstein In other words, there's just as much danger in oversimplifying trading as there is in overcomplicating it. Optimization is finding the right balance. :cool:

:eek: Hear, hear!! :) :) I am a confirmed member of the roll-your-own school of TA. :cool:

So according to your definition, there's no such thing as a trend, because there has never been an infinite sample of data collected even if we add up all of human history. That's ... unique. :D :D

The ker doesn't require an infinite sample. Maybe you should reevaluate your definition of trend.