I couldn't open, this, Gann. It opened horizontally, and there are only A-ZZ posibilities in excel, and....awww, damn, this file is so large, it crashed my STATA 9 Intercooled Software. Kudos. I can't work with just the text file, but I would love the opportunity to see the distribution. I'm not doubting those are your trades, and don't think I'm just looking to say something negative. I'm sorry I gave you that impression. Please do your best to compile into excel for me. It's unusable.
I think it goes all the way to the explanation you choose to ignore. As I've said more times than I feel I need to, but maybe I'm phrasing it differently for others to understand: <b>What you're getting at is covariance stationarity. My pairs model specifically <i>transforms the data</i> into a covariance stationary dataset. Is that clear enough? It can be done in Wealth Lab 4, and Wealth Lab Pro, but the secret requires a few more steps beyond thinking of all the data as a z-score.</b> If you don't believe you can transform non-covariance staionary data into covariance stationary data, you need a lot more training such as Level II of the CFA exam to test and learn from. I was just fortunate in recognizing at Level I what a "standardized measure of mean reversion" means, and spent the next two years trading after that for the first QID QLD system. Tests for normalcy are quantitative, but I guess can require some finesse with the interpretation. I think it has helped both of you to see the distribution as what I was seeing. I see a distribution with kurtosis below 1, which I still believe is the actual value excel places on a normal distribution, and the positive skewness. Qualitatively, just looking at the graphical chart of the profit distribution appears as I have described.
The answer reminds me of the proverbial "Grand Yahoo Number 1 Lecture that summarizes all of finance:" Students walk into a room. They sit down. Teacher makes them wait 5 minutes for their first lecture on theory of investments. The professor walks in, sits at his desk, and says, "Everything is perfectly priced. Any questions?" The irony is so vivid in my mind. My professor was a bit more creative in his abhorrence to trading and non-buy-and-hold investing. His statement went, "If you guys believe you can spend your life as a trader or day trader, you need to do some serious soul searching." I did, and I have found a good balance between the traditional and what might be a new frontier known as Financial Economics.
<b>The Pairs Model</b> I'll try to summarize my initial thoughts before getting to the Performance Summaries. For your pairs model, specifically what I am most excited about is your description. You hold 2-20 days, and while I've only found perfectly correlated pairs to trade on, what is <b>REMARKABLE</b> is that your maximum trade profit was 15%, which is a little short to my own. Even further, your minimum profit or "greatest loss" was 5%. My distribution compared to yours I would say is nearly identical. It actually excites me even more that our maximum profits and losses were essentially identical, and the distribution is consistent for daily trades. average 0.011740944 kurt 15.04170834 skew 2.382459037 MAX 0.1519635 MIN -0.0574064 The kurtosis is quite peaked, and extremely positively skewed. I'm a little miffed about this stat: Win Percentage 0.923761118 If you were to go back to my average trade, you'd see it is more than twice your average trade. I would bet you're combining some intraday procedures as part of a long term weekly trend. I think this may be effecting your performance, but, of course, who can argue with a 92.37% win percentage? It does not appear that it has fat tails, more than that it almost seems like you're pairs model is a bit concerned about "preserving capital." A little loosening of your autostops criteria should more than double your trade profit, but will adversely effect your win percentage in all likelihood. So, that's my post <b> based on only the distribution, what I think.</b> I will get to my thoughts on the performance summary shortly, and I appreciate your patience as I compose my responses. The next post will be about System 1.
average 0.027947917 kurt 23.72214953 skew 4.049081018 MAX 2.896 MIN -0.882 Win Percentage 0.447916667 Are you sure this is <i>your system</i> The max 2.896 is ironically what ArmChair Trader's was, so I'm left to think one of two things: 1) You compiled ArmChair Trader's Distribution for me, leading me to 2) <b>You ARE ArmChairTrader</b>. Ha. Anyway, you're a funny guy, Talon. Here's what I think. The kurtosis and skew doesn't matter to me in this case, because there is such a wide disparity in the maximum and minimum values that it would suggest that there is inherently a lot more risk with this system. It is also evidenced by the win percentage being 44.79%. The only thing that I do know about it, is that the kurtosis is extremely peaked. This distribution with what is to date the highest skew value reported in this thread tells me that your profits are extremely large with this system. Having some four baggers and 100% profits really does add up when your distribution is heavily skewed positively. Sadly, I can always say, and I think anyone else can, too, that having a trade lose 88% of it's value is unacceptable. You need to add a stop around 30-40%. Please do me the favor of updating that and showing the change in the distribution and performance summary. Of course, without knowing MAE and MFE, it's not possible to know with certainty. Maybe we can have that be our next distribution analysis. So, huge profits, sometimes huge losses, but resoundingly positive. <b>Again, I will now analyze only the distribution of System 2, and I have yet to comment about the performance summaries. I appreciate your continued patience. </b> ex post edit: On the second pass I realized it's most likely an options strategy, requiring me to re-assess what I think. I will return to this day after tomorrow. It's a little late, and I was up early, so I'll post back on Friday. Good Trading, Good Luck, Beau Wolinsky
Ok no problem. FYI, this is not an options strategy though. It's a hedged stock strategy... in fact, I'll go ahead and let the cat out of the bag: This system is our actual return trading that S&P Add Delete strategy that I posted in that thread I started several weeks ago. Funny, because that's how you and I got off on the wrong foot in the first place... but here you can see the returns. And yes, they are a bit wild. Please check your numbers... Excel calculates statistics incorrectly and you cannot use Excel for data analysis. I posted correct stats (from Stata) on the back sheet of that workbook... so compare what you're getting with Excel to those. I posted a while back on this issue... I've seen several attempts at system backtests killed by Excel's crappy math. FYI, we actually do some data analysis in Excel, but we have a VBA module with code to do correct stats. I do understand the temptation of using Excel, but resist! Get to the rest when you can. Oh and System 2 is the good one.. especially when you consider the source!
Beau, One can transform anything into a linear model. This is what turns me off to most (if not all of) published econometric models and why I consider it pseudo-science. Non-linearity is a trait inherent in 95% of all natural phenomina - especially in markets. IMO, simply because professors can easily teach a nice linear model, then create nice exam problems for it, is the reason these models are so prevelant in most cirriculums and also the reason why we see so much literature about them. They're easy to underatand and easy to implement. The real world doesn't work that way, most everything is non-linear... I'm suprised you dismiss cauchy sets so easily. I think it would be worth everyones time to understand cauchy sets. In fact, they work very well when applied to volatility modeling - easily beating most autoregressive models, ARCH etc. I have done the work and cauchy sets are by far the most representative of non-linear market dynamic, specifically when modeling volatility. Also, I would think one would need a lot more than three years of system data to determine if the data set you have is normal - quantitatively OR qualitatively. Mike
Cauchy is not something basic. The adage that goes the simplest models work the best is a valid one. Cauchy is beyond what's required to build a viable, profitable trading system, and I also don't see how a distribution can be used to model securities prices. WL says the minimum to validate a system is 1500 trading bars. About 6 years of a backtest can validate any model. Right now I think I'm only over 800 bars, but a lot of the models I built on WL are more than 3 years old now, and the data only went back so far when I created them. The thing about these systems is that I can identify a valid, statistically significant system with both positive expectancy and high probability of future success. That's pretty much all I'm offering, and all my trading is. <i>Here's a model. Here's how I got it to perform, now shall we see what happens when we trade it?</i> That's the stage you have me at, and I will be succesful with these systems.
>Cauchy is not something basic. Well... um... yes it is... I guess it depends on your perspective. >The adage that goes the simplest models work the best is a valid one. True... as simple as it can be without being any more simple is a better way to put it. You're making things a little more simple than they can afford to be if you get my drift? >I also don't see how a distribution can be used to model securities prices. I guess that's not on the curriculum at Centre College, but I can tell you I interview kids from Chicago, NYU, Columbia, etc who do understand what we're talking about here and who understand it very well. Finish your analysis and we'll talk more... I'm trying to decide how best to effectively communicate what I want to say here.