I’m Robert Carver, a former head of fixed income at Man AHL. I specialise in quant trading systems for systematic trading and am now a lecturer and author on this subject.
I’m publishing a new book – the Art and Science of Trading - in in December. You can read about it here https://www.systematicmoney.org/art-and-science-of-trading.
Ask me anything about systematic trading jobs! I’ll be around to answer on September 17th and 18th.
I’m a fresh graduate based in China and want to break into a trading job at Man Group — would you be able to help?
What would you say is the biggest difference or challenge for systematic fixed income trading versus something like equities, which has a longer history of being traded algorithmically? Are there any areas of fixed income which are yet to/unlikely to be infiltraded by algo traders?
Hey Robert. Thank you for this
Just wanted to ask - what do you think the next big trend in the industry is? Firms are stocking up on GPUs etc., but to what end? What is their endgame with them, really?
Hi Robert, thanks for taking the time out to do this. I have 3 questions:
how can quant be used to determine position sizing for trades? Are there simple models that can be used or is this a blend of art and science?
what would you reccomend as a starting point to learn how quant can be used in Asset allocation and tactical asset allocation.
could you recommend interesting reads or subscriptions that help blend quant and investing/ trading?
I don’t have a quant background but am open to learning through self education.
Thanks in advance
In this same vein, what would you say is the lowest hanging fruit for non-technical people to become a self-employed systematic trader? I couldn’t write an algorithm to save my life, but can I just get Claude to do it for me these days?
Hi Robert, appreciate you doing this.
I do electronic FX market-making at a bank; pricing, internalisation, managing inventory risk and optimal hedging. Pretty different from trend/macro systems: little forecasting edge, and much more adverse selection control and risk sizing under uncertainty. Where do you think the “art vs science” split in your book actually shows up in market-making vs directional trading? Same underlying problem in a different wrapper, or genuinely a completely different discipline?
Separately, I’m at the point of my career where the next step will be making VP on the sell-side, I debate if there’s a point to try to progress further or in trying to make the jump to something more research-heavy on the buy-side. Of course, the skills aren’t directly transferable. When you look at people from a MM background wanting to move into systematic trading, what are the main things you tend to see missing?
Hi Robert, thank you for taking the time to do this!
Just wanted to ask - do you think systematic trading is kind of doomed in the world of AI/ML? I struggle to see how even the world’s very toppest quants can compete with a GPT-8 or whatever will come in the next few years.
Ha! Nice try but it’s many years since I worked there.
Biggest difference is bonds are more heterogeneous… most firms have 1 class of stock, but different maturities and coupons for bonds plus weird other stuff like convertible. So the bond market is wide but shallow. As a result it took longer for electronic markets, a prerequisite for serious quant trading, and that lag is still there. A lot of trading is still OTC so higher barriers to entry, and data harder to get. Technically obviously you have higher correlations in the pool of instruments, some very high, which produces problems and opportunities. Mispricing is probably rarer since most investors are pros not crazy retail people. A smaller more identifiable set of factors drive the market. I could go on… I think these difficulties create more potential alpha but there are no patsies to take it from!!!
I hate questions asking me to forecast the industry. There is a good reason why I delegate my forecasting to mechanical models- I’m not very good at it! Yes AI is the current big thing. I think it has it’s uses but they are much more limited than the hype suggests and capturing any meaningful value will be much harder than just buying chips and fitting your own local LLM or anyone could do it!
Position sizing is pure science and that’s a big theme of my new book. The only slight caveat is a quantified measure of confidence can go into sizing positions and that could come from a discretionary process.
For your second and third question I’d also recommend the book as a starting point. I’ve written it for non quants.
It is extremely hard to be a self employed trader, systematic or otherwise. I would really push back against the idea unless you have some special knowledge or skill you could transfer into profitable trading. The simple systematic models I have in my books could potentially provide you with an income but only if you are already pretty wealthy.
Obviously there is some similar crossover knowledge. But probably the biggest difference is you can’t really backtest an MM model. I worked with a lot of physicists in derivatives MM. That was a very natural fit. But I noticed some physicists struggled with evaluating backtests because the role of uncertainty is much bigger. And yes AHL was founded by 3 physicists and I also worked with some very good ones… but I would be similarly concerned about someone coming from an MM background as a fresh PhD physicist because there isn’t some universal model of truth hiding in the backtest.
I disagree. Profits being arbitraged to zero when smarter new entrants arrive has happened in markets since the 1st guy turned up with an HP calculator instead of a slide rule, and probably long before, and yet trading continues to be a lucrative activity for some.
A great way to market the book, thanks