Nimble Market Maker - Monkey Patch and Benchmark/Testing
& some old alpha in the Polymarket API
A number of months ago, I started the nimble market maker alpha series -
We then talked about how the insurgence of crypto and blockchain matching platforms provide small quant teams and even individuals to deploy high sharpe strategies in pseudo-competitive markets due lower barriers and fragmented liquidity.
The nature of these budding platform often have attendant issues in one form or another, including behavioural and performance ones. These issues also often take weeks to months/years to get ironed out, some despite already having been flagged.
Whether it is engineering inaptitude (monolithic stack), prioritisation (these platforms often have a multitude of pressing concerns and distribution objectives), awareness (the issue is not recognised by leadership) or negligence (aware but don’t care), often flaws exist.
In some instances, you can shout at them “THIS ISSUE IS COSTING YOU HUNDREDS OF THOUSANDS” and they will give flaky answers, leaving you to wonder if you should have told them in the first place or taken advantage…
This was a conversation in the past year from [REDACTED] crypto platform, where the reply to my ticket was how much money they had raised!
In most cases, the issue is more subtle, such as performance. Undocumented API endpoints, wrongly documented endpoints, undocumented parameters and so on. These can be rectified to some extent by looking at the developer tools (Chrome F-12) network payloads and observing the message traffic.
In other cases, everything appears as is, and no behavioural flaws are apparent. However, if you are looking to onboard a platform and play a significant role in liquidity provision, it is absolutely necessary for you to understand the in-and-outs of both the functional and behavioural kinks of interfaces presented by the exchange.
For example, a cancel order functionality may be available via these features: