4 ms·
QuantMath: Financial maths library for risk-neutral pricing and risk in Rust
- fancyfredbot 6y agoGreat to see Morgan Stanley allow quants to work on open source projects directly related to their day job. I do not think all banks are as open minded!
- orange_tee 6y agoI am guessing but since this is for pricing and risk, it's a very different game from trading. In this particular situation they prefer to have a reliable piece of software that is shared and used by all. Just guessing.
- klelatti 6y agoCan anyone who is closer to this field suggest what the major options for open source libraries in this area would be? Quantlib? OpenGamma Strata?
- p33p 6y agoThose are the two most well known options, yes.
- 9NRtKyP4 6y agoIt’s still under development but some at Google are writing a quant library in Tensorflow. https://github.com/google/tf-quant-finance https://github.com/google/tf-quant-finance
- klelatti 6y agoThat looks really interesting and under active development - thanks.
- surfsvammel 6y agoInteresting. I will keep an eye on this. Two comments though: trying to handle corporate actions, instruments, pricing, risk as well as settlement in one library might be a bit too ambitious. Maybe it would be better to focus on just the pricing, or the risk part of things? The other comment, is that I would probably split products from the pricing. I looked at one of the product definitions and it seemed to be intertwined with pricing. Products and pricing can both be complex in an of themselves, and they are not necessarily conceptually coupled. Just my two cents.
- Dowwie 6y agoNote the last commit date
- adamnemecek 6y ago8 months ago?
- koolk3ychain 6y agoWhat kind of maths are these libraries based on? Always been curious how math is used in finance to build investment strategies / gauge risk.
- arthurcolle 6y agoUseful reading: https://en.wikipedia.org/wiki/Mathematical_finance https://en.wikipedia.org/wiki/Mathematical_finance https://en.wikipedia.org/wiki/Black%E2%80%93Scholes_model https://en.wikipedia.org/wiki/Black%E2%80%93Scholes_model https://en.wikipedia.org/wiki/Valuation_of_options https://en.wikipedia.org/wiki/Valuation_of_options https://en.wikipedia.org/wiki/Binomial_options_pricing_model https://en.wikipedia.org/wiki/Binomial_options_pricing_model https://en.wikipedia.org/wiki/Replicating_portfolio https://en.wikipedia.org/wiki/Replicating_portfolio https://en.wikipedia.org/wiki/Greeks_(finance) https://en.wikipedia.org/wiki/Greeks_(finance) https://en.wikipedia.org/wiki/Delta_neutral https://en.wikipedia.org/wiki/Delta_neutral
- koolk3ychain 6y agoWow, thanks for putting this together!
- arthurcolle 6y agoAnytime, old sport
- arthurcolle 6y agoNice, this is great. I was recently looking for some QuantLib bindings for Rust and its not really there yet. There's a great book called Implementing QuantLib by Fabio Ballabio that is a really nice introduction to QuantLib for those that are interesting. It is really nice, has lots of term structure data structures built in and some datetime abstractions useful for pricing/risk calcs, and supports lots of financial instruments, like CDS, vanilla/exotic options, bond calculations. Just wonderful.
- twic 6y agoFun, but seems extremely limited. I can't see a way to bootstrap a yield curve for interest rates, which is the first thing i would need to do. The standard model for curves seems to bake in assumptions about day count convention and interpolation: https://github.com/MarcusRainbow/QuantMath/blob/b51ffacc2cfe5b0c35578537341b207f3eb798ed/src/data/curves.rs#L118 https://github.com/MarcusRainbow/QuantMath/blob/b51ffacc2cfe... There is no provision for swaps or futures. Perhaps this is not aimed at rates use cases?