Weekly Algorithm Review: 09/23/2023 to 09/29/2023
Performance Rankings
Sector Neutral: +0.04%
Overall Market: -0.55%
Variable Sector Neutral: -0.96%
Variable Market Neutral: -1.39%
Base Algorithm: -1.61%
Market Neutral: -1.64%
Long Term Portfolio: -2.52%
This was a much stronger week for our systems. Winning by 91 bps is fairly good for the base algorithm. Unfortunately, it was matched by a 197 bps under-performance by the long term portfolio, leaving our performance below the market this week.
While it was a losing week for us overall, this bodes much better for the algorithm than the last few weeks have.
With a difference of this size, a few tickers have become standouts. First off, we managed to avoid much of the trouble with NEE this week by under-allocating into it. Its worst day this week was Wednesday, going down more than 7% intraday. The algorithm only allocated into it all after this, on Thursday and Friday.
Additionally, the tickers the algorithm heavily favored this week performed well. Heavy allocations like LHX, ERIE, and MPE performed fairly well. These were outweighed by losses from other heavy allocations (CVS and MET being the most obvious), but performance was still overall much stronger than the portfolio overall.
The real kingmaker this week (or, king-breaker rather) was the Utilities sector. It under-performed the market by a margin several times greater than any other sector we consider. If were heavily into Utilities this week, you probably under-performed the market by quite a bit. If you weren’t, you probably did fine. The algorithm, as it happens, had nearly 60% less capital in the sector, when compared to the long term portfolio. This is one of the biggest reasons for the algorithm’s outperformance this week.
What’s In The Pipeline
This was a very strong week, especially when compared to other recent weeks for the algorithm. But one strong week does not a comeback make. I’m going to want to see continued out-performance from the algorithm. If we have a strong success this coming week, or just moderate success for the next 2 weeks, I’ll be inclined to remove the hold. I know it’s been frustrating for all of us, but we will not endorse a product without completely believing in it. I don’t like the algorithm hold any more than you all do, but we still feel it’s the correct option. The algorithm is already up ~8% YTD over our portfolios. The strong success we’ve already seen makes me more willing to wait for better results again.
In the meantime, experimentation with updated versions of the algorithm remains in progress. The current experimental version I’m working on has taken several days, and is nearly ready for backtesting. This includes a rework of many of the alpha signals our system uses. If this outperforms, and is released, it will be the largest modification that has been made to the algorithm in the last year.
Additionally, expect some minor updates to TradeBot this week. No new features are being developed at this time, but some usability fixes. TradeBot sometimes sends the same graph multiple times. While not a major issue, this does negatively affect readability. I’ll be looking into ways we can stop the LLM from doing this, or adding some outside filtering if necessary.
Misc. Data For The Week
