Algorithmic Week In Review: 08/21/2022
Current Algorithms - Recent Performance
My main focus this week has been how effective our modeling is, and how consistent our hedging. All in all, I’m happy with both, performance-wise.
Results have largely been within expectations. Even on days our hedgers underperform, they usually outperform the long portfolio had we held the whole thing and hedged the same way. This leads me to conclude that our modeling and hedging systems work together well. We still forecast some market troubles in the near future, so this is all great news. Our biggest priority is pursuit of projects that will help us to weather that.
What’s In The Pipeline - Short Term
Renaming our systems. We started naming our algorithms after songs when we were experimenting with several trying to do the same thing (back when we first released Kokomo, our second market neutral system, we thought it would be too confusing to name it “Market Neutral 2” or something like that). This is no longer a problem we’re facing. While we have 2 sector hedgers (and are about to release a second market hedger to beta), only 1 is public. The other (FeelGoodInc) is in the beta channel only. For this reason, we feel comfortable renaming our systems. Expect less “GoodVibrations” and more “Sector Neutral” starting this week.
A New Market Neutral Algorithm. As alluded to in the last point, we have a new market-neutral system in development. It works the same way FeelGoodInc works in relation to GoodVibrations. I don’t want to go into technicals here, but the short version is more accurate modeling of technical signals should lead to better performance. Backtesting has gone quite well. I’m expecting to release this in the beta channel starting Monday night.
Upgrades To The News Bot. We upgraded our news bot earlier this week. If you regularly check it, you’ve probably noticed that it’s much more consistent with its ratings now.
What’s In The Pipeline - A Little Longer Term
Upgrading Our Trading Bot: This, unfortunately, has continued to give us trouble. We are still looking into why it isn’t staying online consistently.
Replicating Congressional Stockholdings: This one is . . . interesting. We are able to get this data, as well as historical records. So we can create algorithmic reports from this, as well as backtest to estimate how effective it is. The problem is that our historical data only includes official disclosures, which can be made up to 45 days after the transaction. Before we can do this, we want to construct a consistent formula to rank a senator as a trader from their disclosures. We also need to know if the stock is done growing. If a senator buys a stock they know is going up in 3 weeks, and discloses it 4 weeks later, our system will have to recognize that it’s done climbing. These 2 formulas are our big road block right now. I wouldn’t expect this in the near future, but keep an eye out for developments.
Crypto Portfolios: We’d like to release this publication as an option for users. Maybe a “Crypto Market Neutral” that hedges against the crypto market as a whole by shorting some popular coins. No testing on this yet, just something we’re looking into right now. If I never bring it up again, it probably either got veto’d or just didn’t do well in preliminary testing.
Automated Trading: This is still a goal, and we’re contact our broker this week to see what our options are. Given the lack of stability from our trading bot, I would want a very resilient way to make this happen. If we can do this consistently, we can start doing more advanced portfolio strategies.
Tomorrow’s Outlook
The full algorithm reports will be published tomorrow morning, once Allen has had a chance to vet its recommendations. In the meantime, here are our tentative exposures for the trading day tomorrow:
That’s all for tonight. Good luck this week everyone, let’s see some big wins.
-Asher
