Good afternoon from Puerto Rico. We’ve had a streak of amazing weather, and today is beautiful again. I remember when I first moved to Los Angeles from Chicago and thought that winter couldn’t be any better. Being able to comfortably go for a night swim or lounge outside without a jacket in the winter here is awesome though.
My week was pretty decent. I could/should have eaten more cleanly, and I got roped into more drinks than I wanted Saturday evening, but my workouts were good. I started tracking my exercise routine again, and I add just one rep to one set each week, but that cumulative increase works well. I can see the progress, I automatically push myself to hit my targets without thinking, and I don’t get burnt out from going 100% every week. Some training sessions I have to push to my max to hit my incremental new targets, but I usually have more in me, and it helps me mentally be confident that I can improve at a pace in line with increased weights/reps. I need to tighten up the diet this week though. I think yo-yoing with a super-strict diet, and semi-strict is more realistic/better than 100% strict and falling to pieces. My main culprit has been pretty simple: I bought a ton of skirt steaks, I keep cooking them badly (I’ve tried a few different methods, but none have worked well so far) in the sense that they are like chewing leather. I get bored of trying to eat them, end the week a bit starving, and easily digress to bad stuff on the weekend in a hunt to get calories. I need to stick with more staples that I can prepare well (ribeye, lamb, eggs, burgers) and buy new meats in small quantities until I know I can prepare them in a way I’m excited to eat, and not stare at 4 lbs of grizzled chewy beef in my refrigerator that I feel I have to force myself to eat.
I was planning to get a new data provider for stock trading research this week, but I ended up working on a few other things that I was antsy to try. The first was building some machine learning models to predict when the S&P 500 went up too much compared to the Nasdaq, and vice versa. I’ve been thinking about this for a while, and would like to trade all of the financial indexes against each other (SPX, Nasdaq, Dow, Russell), but not shockingly in one of the largest and most competitive markets, it was very hard to find any quantifiable edge. I had limited data, and I didn’t spend a ton of time tweaking the models, but the base version of my stock trading machine learning models didn’t seem to have much juice with the big indexes. It was gnawing at me though, and I had to know if I was putting all of this effort into individual stocks that could potentially work just as well with the big indexes.
The other thing I did this week was personally trade the open and the close of the stock market. All of the work I normally do is with data, and if I’m not coding a program to make trades for me, I’m coding something that tells me exactly when and what to do if I have to click the mouse myself. I’m never watching screens and buying/selling based on what I’m seeing, maybe a few times a year, and that’s only to do one overriding trade, not trading back and forth. I’ve been listening to podcasts from all of these “day traders” though, who are frantically buying and selling throughout the day when they have setups and other things going on. It’s what you would picture normal people in trading actually do. I decided to try it to help me learn something, and cocky me figured I would make some money too. I did this knowingly that this was more of a brainstorming exercise that would help me generate new ideas than replace my daily work. It’s still funny that I went at it with no plan, strategy, or set of instructions. It was more of just looking at stocks that were cratering, and trying to buy them before they made a recovery.
I clicked with my mouse on my laptop, with my slow internet, trying to buy stocks that had bottomed out, but were about to go up a lot. The first few days I was very conservative, waiting for stocks to go down a lot before pouncing. They all went down, but not down enough for me to make a trade. Then I started getting more aggressive and actually bought some stocks. Some went up a little, then stopped going up, and I sold for about even. The other ones just continued to go down, and I sold them for a loss. So basically I either broke even on my trades, or lost money. I traded very small quantities, so the money wasn’t a big deal ( I think I lost around $80). After two exhausting days of this (it’s 100% focus and tiring when you don’t really know what you’re doing) I realized something pretty funny… I was watching all of these stocks that I felt confident were going to keep crashing, and they did, yet I was only trying to buy them! In the markets you can usually bet in either direction, and sometimes betting down is easier. The problem shorting (betting down) with volatile stocks is that they can easily go up a lot, and you can theoretically lose infinite money, but you can just bet smaller/take losses if it gets bad. I consider myself open to any strategy/idea in trading, but I’ve been trying to avoid bets where you can make 4% but lose 100%. I know other people don’t want to make these bets either, so I believe that’s part of the reason they can be profitable.
I remembered some datasets that I built a few months ago ( it feels like years), and they coincidentally had all of the data that I needed to filter and explore my theories. I had a few naive ideas, but working with 15 years of data I was able to find some simple rules about what stocks were likely to keep going down, and some hints at which ones would reverse. I still have plenty of optimizations to test, but I’m very excited to get it running and automated this coming week! I have my once-per-day trading strategy with machine learning that I’m already running, and now I’m planning to add the down-momentum algorithm, plus one strategy that seeks to buy those names in some rare circumstances when they might bounce and go up quickly.
Exciting stuff if it works, as usual. This should provide a steady stream of trades though, so unlike my daily trades (that sometimes doesn’t have any trades that meet the criteria), I’ll have a lot of action and a stronger idea of how it’s performing in the near future.
I wanted to end this post with a passage from the book “Madness, Rack, and Honey” by Mary Ruefle. I don’t remember how I heard of this book or why I put it on my Amazon wishlist, but my sister bought it for me for Christmas, and I love getting little doses of it. The book is a supplement for her lectures about poetry, but I have no clue how to really describe it. It’s one of those books that goes on about random things in life/art/writing, and for my taste, it’s really well done. I like to start my day with a chapter lately. This is from the chapter “Someone Reading a Book”:
“We are all one question, and the best answer seems to be love– a connection between things. This arcane bit of knowledge is respoken every day into the ears of readers of great books, and also appears to perpetually slip under a carpet, utterly forgotten. In one sense, reading is a great waste of time. In another sense, it is a great extension of time, a way for one person to live a thousand and one lives in a single life span, to watch the great impersonal universe at work again and again, to watch the great personal psyche spar with it, to suffer affliction and weakness and injury, to die and watch those you love die, until the very dizziness of it all becomes a source of compassion for ourselves…”
Andrew Manser 