Tuesday, August 24, 2010

The Big Stock Post

Been telling M. Lydon I was going to do this for a while, and I'm finally doing it.

My plan was to come up with a quick scoring "algorithm" for the fundamental basis of stock value & potential, using 10 commonly-looked at financial variables/metrics that are supposedly correlated with the measurement of a company's "success." (by "success" I mean "stock price increase"). There are many, many more variables that professionals use to analyze the worth and value of a stock, but some analysts use only price charts to decide when to buy and sell stocks.

100 Stock Analysis - stocks from Dow Jones Industrial Average, S&P500, Fortune500.

I wanted to make an average "value score" of established, big stocks so I could compare future stocks I wanted to analyze to it, in search of promising companies with potential. I wanted the average score for 100 certifiably good companies to be "100" for convenience, but it fluctuates with the overall market. Here's the list of the 10 components to generate the score ("Value Metric PE", column AA), in order of their weighted importance:

EPS Change (year over year)
EPS Rank (Investors Business Daily)
Estimated Annual Growth
Profit Margin
Sales Per Share
Ideal Growth to P/E Ratio
Ideal Price to Sales Ratio
Insider Ownership
Quick Ratio
Price Ratio to 52-Week High

As I've also recently been working on the charting part of it, I've been trying to implement a combo of what's considered "fundamental" and "technical" analysis -- buying solid stocks at critical moments in time (e.g., at the beginning of an uptrend after a recent downtrend). 
Technical Event Name (worksheet link)

ave. 10-day gain ave. beat 10-day market
Engulfing Line (Bullish) Average
1.93%
1.25%
Exhaustion Bar (Bullish) Average
4.45%
7.42%
Gravestone (Bullish) Average
0.72%
-1.99%
Hammer Average
1.60%
0.09%
Inside Bar (Bullish) Average
-0.27%
-0.61%
Inverted Hammer Average
-0.74%
-1.79%
Island Bottom Average
-3.06%
-2.81%
Outside Bar (Bullish) Average
2.56%
0.84%
Two Bar Reversal (Bullish) Average
-0.68%
-1.23%
Grand Average
0.99%
0.07%

Some "Technical Events" are better than others, but they don't necessarily mean anything without "confirmational" signals, such as increases in trading volume, momentum and other measurements of interest in the stock.

But what I've come to learn is, some days all of that preparation doesn't mean jack shit. The market's just gonna do what its gonna do, the bears are going to rip through town and take everybody down with them. I don't know how almost all stocks could be interconnected but they pretty much are and from what I hear this is more of a recent phenomenon... The only way to avoid dependency on the market is by buying low Beta stocks (low market correlation), but that doesn't necessarily mean that they are going to be a good stock.

Then there's the shorts, the people who get a whiff of fear and cause major price drops by selling institutional shares on loan in hopes they will go down further in the future (assholes). Then there's the old billionaire or millions of other aging investors who have decided they've had enough and are cashing all their shit out at the same time.. It looks like it might be a good time in history to learn the art of short-selling, but knowing my luck shit will reverse course right after I start doing that...

So, every weeknight I have a list of stocks emailed to me that happened to trigger a Technical Event during the day. I look through them, first looking at their current trends (I'm only buying shit in longterm uptrends right now), what happened to the stock that day, and then I generate value scores for the more promising-looking ones, sticking the higher-scorers in a Google portfolio so I can track them. Then when the market opens the next day (3:30 am here), I watch their initial progress, and the overall progress of the market to make sure its not gonna have a decidedly shit day today. If things look decent, I buy the stock from the Google portfolio list that has accelerated the most so far that day.

Then, when my shits fall anywhere from 5-30%, I sell them and buy new ones... And that's how you get reamed in the stock market!!

No, but really.

I've actually only been implementing this entire strategy for a week or so now. I have a rigorous set of guidelines laid out, and its tempting to not follow them or cut corners, but I am gaining disciprine and we'll see what happens from here on out.

If you made it this far and am interested in what I currently have:

ArQule, Inc (ARQL)*
Corinthian Colleges, Inc. (COCO)*
Zhongpin Inc. (HOGS)
HMS Holdings Corp. (HMSY)
ProShares UltraShort Yen (YCS)*

I also wanted to buy the Chinese stock Harbin Electric, Inc., but TradeKing won't let me put the order through without calling them. Its for your own protection, they say.. Maybe its because about 25% of this company's stock is held by short-sellers, which is remarkably high, meaning people think the demand for electric goods in China is gonna tank or something.

Matt Lydon? Are my value metric weights scored copasetically?

What do you hold right now, out of curiosity?


*bought 8/25

12 comments:

Huge Larry said...

My understanding is that the stock market is a chaotic realm and quite difficult to tame. I fear it because I know it has the ability to rape my cash reserves.

Mondu said...

Good luck with the numerology bro. Hopefull you can make me eat my words.

Mondu said...

Well, its not quite numerology. I think an algorithm like this could give a shallow indication of a stocks 'average success value', but I'm skeptical about the reliability of this data. We need a rational basis to justify the risk of our investments. This can only come from fundamental analysis and basic conceptual and logic reasoning. Once we have constructed a 'meaning context', focused technical analysis can help us map its terrain. But that context must come first. I don't think the global approach will work, but if it does I'll buy you a beer.

cyrusfx said...

"We need a rational basis to justify the risk of our investments. This can only come from fundamental analysis and basic conceptual and logic reasoning. "

Yeah, that's what I was going for...

What fundamental analysis do you think I'm lacking? What specifically is illogical about my strategy?

I'm willing to learn more about options, never really thought about it.

Mondu said...

Historically prior metric values do not cause future metric values. Rather, numerical data is a manifestation of underlying causes. We need to find the real causes behind the numbers before we make investments. Patterns are discernable within the numbers, but they are shallow without the meaning context I described. By trying to funnel market data into a general purpose algorithm your setting yourself up for failure. You gotta get over the mathematical bias and focus on the relevant issues.

There are no metrics for the following values, though I argue they are the most important sort:

1) Company's capacity to maintain efficient management structure.

2) Average brilliance of company's newly hired engineers.

3) Functional superiority of companies new products as weighed against competing companies new products.

4) Likelyhood of a company to secure sufficient gov't contracts.

5) Measure of board members commitment to productivity and confidence in the companies business model.

6) Average movement of large investors in/out of companies relevant market.

7) Amount and rapidity of scientific research that will make a big seller obsolete.

And so on...

The truly relevant factors are difficult to shoehorn into numbers. Once these factors are properly situated in context, then technica analyis might be able to give us a bit more edge.

Honestly man I just dont want to lose too much money because you think you can use math to predict the stock market. Just trying to give you a reality check.

cyrusfx said...

"Historically prior metric values do not cause future metric values."

I'm not predicting metrics. Most of the components of my scoring system are all time-proven indicators of the financial success and robustness of a company. Various teams of analysts have looked back at a hundred years of stock data for common metrics among companies that indicate stock growth and have concluded that the fastest-growing companies tend to have these things in common.

The scoring system is simply one of many tools I use to help me decide what to purchase -- it just indicates the potential of a company - and I don't solely rely on it.

"There are no metrics for the following values,"

Actually there are for some of that information (investor movement is followed by volume for example), but some of it is simply unknowable or unmeasurable.

The rest of the information that is knowable and relevant is already built into the stock price and its current movement, because there are already hundreds of people with millions of dollars that are up to date on whatever possible research is currently obtainable.

Since human behavior tends to follow the same patterns over time, Technical Analysis can shed light on when a sustainable trend is happening or is about to reverse course.

Saying I'm "using math to predict the stock market" is true. We also use math to talk, to drive, to play Xbox, and to roll joints.

Did you know that 75% of all trades done on the U.S. markets are automated? That means 75% of all trades are entirely math-driven.

So of course you can use math to predict the stock market. People have been doing it since the 1600s.

Mondu said...

Ok, so we have some data before us.

There are two ways to understand this data. We can find a surface pattern in the sequence of numbers (ie connect the dots, regression analysis, etc.), or we can uncover the causes of the values (ie. tsunami lowered values here, new tech raised values there, etc.). Patterns are easy to find. But genuine causes are not. My point is that causes give us the power and justification we need to invest. Math is useful of course, but only within and after the establshment of the proper context.

I think these 'non mathematical' factors are not as unknowable as you think. Also, I don't believe they are always already integrated into the market price (sometimes they are). The trick is to know something that not many others know yet. Or better, the trick is to recognize as relevant something others dont recognize as relevant yet.

Mondu said...

The abstract, conceptual reasoning power of mind far outstrips the capacities of sequential, algorithmic computation. But thats just my bias.

cyrusfx said...

I am coming from the standpoint that it would be narcissistic of me to assume I could outsmart thousands of professional market analysts, hundreds of which have double-PhD.s in mathematics and economics or whatever combination the Wall Street institutions hired them for, who have combined investment capital in the billions or trillions and decades of experience each.

What I can do is see what they're doing and move in right afterwards.

Your approach to stock buying might work for some, but you're more offering condescension instead of a real plan I am able to put into motion.

What have you found to work out of personal experience, as you seem to be dismissive or hundreds of years of tried-and-true methods? What stock is a good example of what you're talking about?

The reason why I posted this is because I know you guys are smart and I'm trying to get some constructive criticism that I can incorporate into my model.

If you think the foundation of the model itself is bogus, you have to be more specific as to why when I've refuted your points and honestly don't understand how to incorporate your last couple of paragraphs into realtime action.

marchhare14 said...

I've seen Pi. There's a pattern there, man, you just need to discover it using a superpowered microchip and the Kabbalah.

cyrusfx said...

Goddammit Josh, I'm not using no fucking Kabbalah.

I'm using the fucking Torah.

Cmon be realistic, now.

cyrusfx said...

i hope that came across as self-evidenting a "j/k"..