Why Apex Predator?

So, why should you bother reading anything I have to say on this topic.  Apart from the fact that I’ve been modeling options, in one form or another, for over 30 years.  The short answer is: because I’m getting pretty good results trading them – which is where the rubber meets the road.  It’s a short track record, to be sure, but it’s been very consistent, and it’s under continuous improvement as the model and trading interact to improve my understanding.  Here are some numbers to put meat on the bones.

xxx

As of the end of August 2024, I’m at 21.3% realized gain from option trading and 23.2% income return, each on an annualized basis. My strategy to better utilize / get rid of low juice deadwood in the portfolio, discussed just below, is already paying off. There’s been a slowdown in trading opportunities the past few months, versus the awesome first few months of the year, but all in all, not bad.

* Income return is dividends plus realized gain on options, calculated using Book Value (purchase cost) as the base.

** Portfolio return is calculated using market value as the base; I adjust market value return for withdrawals using the Dietz algorithm, trying to put my portfolio results on an apples-to-apples comparison basis with the SPY, which has no withdrawals.

***Option trading commenced in August 2021; prior to that I was trading mostly dividend growth stocks using the MACD. I was using information obtained from Sure Dividend to select my stocks.

xxx

Note: 2021 and 2022 performance results are not ‘clean’.   During this period, I was still ridding the portfolio of legacy holdings – old CEFs and BDCs that had not done well, selling them at a loss.  To get an unpolluted measure of the results of the option trading, I’ve made what I think are reasonable adjustments to the dividends, realized losses, book value and the market value of the portfolio to exclude the effects of the legacies.   As these legacy holdings were turned to cash, the portfolio base was augmented accordingly for purposes of return calculation.  2023 is my first clean year with no legacy holdings to be rid of:  all gains and losses are fully reflected in the results as recorded by my broker, Fidelity.  All my trading is done in IRAs; hence, there are no tax inefficiencies or consequences to my trading activity.

Last note: I took an inventory at the end of the year 2023, and determined that about 15% of the portfolio was effectively dead to option trading (though these stocks provide a good dividend in most cases) and another 15% was in relatively non-juicy names or where the current price is relatively far below the purchase cost.   The 17.3% income return shown in the table above was calculated on the entire base, including these sub-optimal holdings.  This means that the active part of the portfolio was generating well over 20%.  

Now there’s always going to be some dead money, unless you’re ruthless about selling these off.   And if you sell them off, the realized losses will offset realized gain.  Either way there’s a drag on results.  The trick is  minimizing this drag. The model suggests that ruthlessly clearing out dead money is the right thing to do, the optimal long-term thing to do.   Right or wrong, in my trading I tend to take a more gradual approach to clearing out dead money.   My goal in 2024 has been to align myself more with what the model is telling me: clear out these underperformers and redeploy the cash.   I’ve sold calls on many of them, garnered some realized gain, and sold off several others – reducing the dead weight by about a third at very low net cost (drag on this year’s performance). More to come as the market reaches new highs.

xxx

Background

Trading stocks using the MACD technical indicator turns out to be a pretty good business – not as good as option trading, but pretty good. I call this trading approach volatility capture.   Here, I extend the volatility capture trading approach to the writing of options.   It turns out to be a very natural, elegant, and lucrative extension.   The combination of option writing, which takes advantage of time (aka theta) decay, and using the MACD to trigger the trade, is both immensely profitable and low risk versus cash market trading.  After developing this approach and successfully trading this way for the better part of a year, I thought it would be instructive to replicate it in a software program.  [That was three years ago, and my trading and model/code continue to evolve and instruct one another].  My language of choice is C, the old workhorse of the investment industry.   I saw two potential benefits to coding up a trading model, apart from the intellectual challenge and leaving something of my trading, modeling, and coding expertise to posterity¹: one, it would allow me to rigorously test various trading parameters to see what combination gave the best results over several market cycles, and two, with a modest further extension, I could implement a trade alert system that would allow me to partially automate the trading process.  This, in turn, would allow me to spend more time by the pool and less in my office.  I have a nice office, but it’s hard to beat the pool.  I don’t plan to fully automate the model to do the actual trades, at least not at this time.  But never say never.  This present work will detail the option trading process, the modeling of that process, and the software code implementing the model. 

Before we begin, it will be helpful to define algorithmic trading and to be clear on what the model described herein does and does not do.   One definition I like is: using software code to execute trades on a trader’s behalf when certain conditions are met.   You might call this is a strong form of algorithmic trading: everything has been turned over to the computer.  That’s not what this website is about, though the content of this website is a foundational part of such an algorithmic trading model.  A very good, more general, definition is this: algorithmic trading is making trading decisions based on pre-set rules that are programmed into a computer.   Notice the flexibility of this definition: who or what is making the trading decision based on the pre-set rules is not specified.   You could readily interpret it as meaning that the model / computer informs the trade decision, with the trader making the ultimate call.   Informing the trade decision could come as specific trade alerts, something we will describe in this website (though the model doesn’t do it yet).   In its weakest form, it could mean that the model / computer informs your trading by virtue of extensive back testing of rules to help you decide how best to trade.  Here’s that wording: algorithmic trading is the trader making trading decisions based on rules that are determined by a model / computer.   That’s the current situation.  Exactly how a model / computer might determine these rules, specifically in the case of option trading, is the subject of this website.   I will also discuss a straightforward extension to have the current model also provide trade alerts.    

1. I’m reminded of a quote by the early 20th century English mathematician G.H. Hardy: “What we do may be small, but it has a certain character of permanence, and to have produced anything of the slightest permanent interest, whether it be a copy of verses or a geometrical theorem (or maybe even an algorithmic option trading model), is to have done something utterly beyond the powers of the vast majority of men”. 

xxx

Next in sequence: Options & Realized Gain: The Basics

Back to Home