Weekly Algorithm Review: 05/20/2023 to 05/26/2023

Performance Rankings

1. Variable Market Neutral: +0.59%

2. Market Neutral: +0.53%

3. Base Algorithm: +0.52%

4. Overall Market: +0.09%

5. Variable Sector Neutral: -0.75%

6. Long Term Portfolio*: -0.87%

7. Sector Neutral: -1.0%

*Hypothetical performance, if we had used the new bullish portfolio all week. In actuality, we were using our old defensive portfolio until Monday night

Before we get into analysis, an announcement. We have switched to a new long term portfolio. This one is designed to be much more bullish and long-market.

As you can see, our new portfolio is more interested in traditionally less defensive sectors - tech, consumer cyclical, industrials, communication services. Likewise, defensive plays like consumer defensive and healthcare have been de-emphasized.

I want to state that, in the long run, we are still bearish this year. However, given recent uptrends in the market, we are temporarily going more bullish with our long-term portfolio.

I also want to note that, looking only at Tuesday-Friday - the days in which we actually used the new portfolio - our intraday returns are:

• Base Algorithm: +0.75%

• Long Term Portfolio: +0.08%

• The Market: -0.34%

For the most part, I’d say the transition has gone well. For the near future, I would take the algorithm’s recommendations with a larger pinch of salt than usual. I want to make sure the algorithm can trade with the new portfolio effectively. Things look good so far, but 4 trading days is a small sample size.

On Earnings

As everyone saw this week, NVDA had some incredible earnings and climbed 25% after hours. We didn’t see any of this, because we weren’t holding anything. This may be controversial, but we consider this to be a feature, rather than a bug. We have 2 policies behind this:

1. We don’t play earnings. The algorithm has a built-in hard rule to not recommend anything around its earnings date (2 days before, 1 day after is typical for an AH report, but we can extend these in outstanding circumstances). This is because the algorithm chooses its positions exclusively based on technical signals. On a day a ticker has an earnings report, most of its price movement will have nothing to do with technical signals. This is the same reason we’re cautious about our systems on FOMC days and related.

2. We don’t hold anything when the market is closed. Simply put, we haven’t found an effective way to model after-hours movement. Volume is much lower, making things far harder to predict. We prefer not to subject ourselves to the extra volatility, as we aren’t sure we can make enough P&L to justify it.

We have considered changing #2, and simulating a hold of the long term portfolio after hours. But even if we had done this, we would have made just 25 bps from NVDA’s earnings. That would have been a great boost, but the extra exposure to earnings and after hours movement just isn’t worth it to us. At this time, we have no plans on reverting either rule.

What’s In The Pipeline?

The options update to our grapher remains in the works. It is currently shelved, due to work on another side project taking precedence. I will give a further update when we resume it.

Misc. Data For The Week