Ok, so now I’m back to theory, but this is just because I had an urge to write during this rather blah week in the US markets. So far I think this consolidation of the S&P and NASDAQ looks pretty bullish and banal, and the VIX isn’t spiking so hard, yet. Thus I’m pretty sure we will break from this bull-flag soon. But predictions don’t matter, right? What does matter? See below!

Beta Go Fish

Lately, I’ve been posting my signals on separate page on this blog. For now, I’m ranking them by beta, with a certain lookback period. It seems to do a decent job, but nothing statistically significant from other methods I’ve tried, to be honest. The reason I chose this is because ranking volatility, especially in relationship against the overall market, makes sense to me. If I choose a stock that meets my uptrending criteria, and I assume this trend has momentum and will hold, then I should be favouring stocks that have higher volatility than the average market to beat it. That is, if the market moves at all, the stock will respond even more strongly. If the supposed uptrend holds, it should respond in my favourable direction (I only go long stocks).

So why choose beta? Because I can come up with a hypothesis that makes sense as to why it would work, and is not some inexplicable data mining result with no rationale behind it. I would like to further this and calculate beta of a stock against industry and sector beta, and should be a fun Amibroker mini-project in the future. I’ve heard in many places that 75% of a stock’s move is driven by market direction, and that sector and industry direction makes up the bulk of the remaining moving power. Idiosyncratic risk is not that great, unless during earnings, bankruptcy or big events (the N in CAN SLIM).

Why even rank? Well with any systematic strategy without a pre-determined watch-list, it basically acts as a further filter on a dynamic list of potential stock purchases. The problem is, it usually contains more ideas than I have money to spend. Therefore ranking helps determine which one to buy first. Systematically ranking stocks avoids any discretionary mistakes due to our inherent biases, which more often than not leads to losses.

Normalizing Volatility

Next step is the process of buying a stock. How many shares do we buy once we’ve ranked them all? Some people just apportion an equal parts of their portfolio to each stock, e.g. 10 stocks, 10% weighting. This is fine and dandy, but if your stock picks range from sluggish blue chip stalwarts to the newest and still profit-less biotech going through FDA approval stages, you’ll notice the more volatile one levering your overall portfolio returns more than the other. This amplifies the risk of one stock over the other, which may lead to you having a stock that dominates the others and overly exposes you to its unique risk.

This is where the concept of position sizing comes in. First you need a stop-loss. The stop loss is a price at which you accept your stock is losing money and sell to protect yourself. You can measure it using your eyes with a chart, standard deviation, Average True Range, whatever.

Now instead of sizing your stock purchase to a portion of your total portfolio, you should size your buy such that the monetary loss from a stock hitting your stop-loss is equivalent to a small percentage of your portfolio. Most traders who use this concept will bet around anywhere between 0.5% to 3% of their portfolio per stock, with most going with 1%.

Example for a 1% bet:
Portfolio value = 100,000$
1% of portfolio = 1,000$
Stock price = 100$
Stop-loss chosen = 92$

Therefore, a loss of 8$ in share price should equal 1,000$ loss in your overall your portfolio

Now based on this, how many shares do you buy? Just use this simple formula:
1% portfolio/(Stock price – stop-loss) = shares (rounded down)

Or even better, simply use Michael G Lamothe‘s Position Size Calculator from his great site ChartYourTrade.com.

So to finalize the example if you choose to do the math yourself:
1,000$/(100$ – 92$) = 125 shares = 12,500$ total value or 12.5% of the portfolio

Check: At the stop-loss, 92$ x 125 shares = 11,500$ or 1000$ less, which is the 1% risk we were targeting.

If you choose your stop-losses wisely, you will notice that more volatile stocks have much wider stops than the more sluggish stocks. From the formula, this means you will buy less shares as the difference between stock price and stop-loss widens. This reduces the impact of a volatile stock on your portfolio, effectively normalizing its force with the other portfolio members. Now you have a properly risk-managed portfolio!

Why 1%?

As you lose money in a drawdown, your portfolio will shrink, meaning the value of 1% of that portfolio will shrink, and thus the number of shares you buy per stock. Therefore using this rule automatically dials down the size of your bets as you lose money, and increases the bets when your portfolio is on an upswing. This is exactly what you should be doing to reduce risk, and I hope to quantify and show you someday. Easy, right!?

Also, 1% is a good number statistically, because this means you can lose 69 times in a row prior to losing half your portfolio! No one is that bad at trading stocks, not even a monkey throwing darts randomly, and in fact several people have proven this. Using effective money management such as position sizing, while buying random stocks effectively causes the portfolio to break-even. So before you pick any stocks at all, you’re breaking even. Now all you need to do is be slightly clever and you’re creating a profit!

Is this too little risk? Not really, just ask the famed Turtle Traders, or most commodities traders, or most talented stock traders. Google position-sizing, you will see, they all made major money in the market. Some risk way less!

Other Risk Management Methods

Beyond simple position sizing here are a few more ideas I may expand on later:

• Profit expansion: You will notice this metric in my portfolio. As a stock becomes a winner, its daily return will have a bigger effect on your portfolio. I.e. if you buy a stock at 100$, a 1% daily return makes you one dollar. If the stock over time reaches 200$, now a 1% move makes you 2$. So while the daily return is 1% at that point, you actually gain 2% from your buy price. Your risk from daily returns has essentially doubled, increasing your heat (how much total risk you have in your portfolio). Some people choose to sell at certain profit points to decrease this heat while staying in the game.
• Management of heat: Managing overall heat usually based on the above tactic or other clever rules, helps keep your risk down. Let’s say you only want 10% of your portfolio at risk at any one time. That means if all your stocks hit their stop the same day, you will lose at most that 10%, unless they gap down hard. What you set that max heat to may lead you to looking out for the:
• Kelly criterion: This is something you may have heard from those famous blackjack card counters. This formula actually defines your maximum heat for you based on portfolio performance. You can also bet “half-Kellys” or “quarter-Kellys” so that you are still optimizing heat systematically but at more comfortable levels.

That’s it for now! Take care,
Mike

Follow me on:
@rocketpower (Twitter)
@rocketpower (Stocktwits)