Here are some general directions:
Alternative Risk Premia
The ARP, or "smart beta," space has gained a lot of tractions over the past few years. These are rule-based strategies that provide systematic exposures to risk factors that have historically generated positive excess returns. Some of the best-known factors are, of course, trend, value, carry, etc. These strategies are well-documented. I recommend reviewing AQR's research library for some of the best practitioner perspectives I've seen. In particular, Style Investing in Fixed Income and A Half Century of Macro Momentum are two excellent pieces that definitely involve rates products. (Disclaimer: I personally am not a fan of ARP...)
Relative Value Strategies
- I'm inclined to think the books you referenced are sufficient – the ingredients are there and just need to be applied in a systematic fashion. Consider a curve fitting exercise (covered in the first book you cited). You can feasibly construct a portfolio that's long the cheapest bonds based on the model and short the most expensive bonds (after accounting for liquidity, funding, etc. and making sure that the portfolio is market neutral) and see how the strategy behaves over time. I have to admit that I haven't seen a lot of published work on this, so it'll be great to see what other community members come up with. But off the top of my head, I remember Citi's primer for their fitted curve model performing just such a backtest.
- A good reference is Bruce Tuckman's Fixed Income Securities, which introduces more sophisticated models that are actually used in the real life in its 3rd edition (although a full backtest was not conducted in the book). Quantitative Analysis, Derivatives Modeling, and Trading Strategies is not a very widely known book, but has a lot of hidden gems in it.
Directional Strategies
- (Mostly) fundamentals-based systematic strats have a long history. I usually reference Ilmanen's Forecasting U.S. Bond Returns. It is decidedly old-school, but it's also simple, intuitive, and very educational. You shouldn't expect it to work out of the box, but it's a surprisingly useful starting point.
- Technical-based systematic strats are also prevalent. Since I don't do much technicals work, I'm not qualified to name classical titles. One piece does come to mind, which is "Predicting Near-Term Performance of the US Treasury Market Using Neural Networks," published by Salomon back in 2001. It introduced a systematic, technical-based directional strat that used machine learning techniques at an era when most people had never heard of ML; it was also live traded.
