1/ For quants just starting out, what specific strategies should you run? Cast a wide net when exploring, but quickly focus in when you have a lead. The zero to one moment of live pnl should be your first goal. A thread on some general directions to get you there 🧵
2/ First up, making vs taking. If you're looking into HFT, this is the first fork in the road. This big question boils down to what you want to work on. Making: Infra quirks, live experimentation, game theory Taking: More statistical study and modeling of market data
3/ Making is also relatively speaking more engineering heavy. When shit hits the fan, if you're taking you can worst case just decide not to trade. However, taking is more winner-take-all. Adverse selection for competitors can be your advantage with some making strategies
4/ Note that making and taking both involve all of these features. The dichotomies above are only what each style emphasizes more. Also, if you're looking at medium frequency or longer, just take liquidity to start. Otherwise you're probably prematurely optimizing. See here:
5/ Next: single symbol vs portfolio strategies (e.g. arb, funding). Again, follow your interest here: Portfolio: A lot of the edge is dealing with risk and edge cases robustly Single symbol: relatively cleaner, risk is clearer, but as a result more competitive
6/ For example if doing spot-futures arb and prices move a lot, other strategies could be sidelined as they hit risk limits and struggle to manage collateral. Single symbol there always be someone trading single-venue around the fair price given all available info.
7/ Which exchange? Pick one of the extremes here. Either Binance and just forget about everything else, in which case you are playing the alpha generation game. Or pick some long tail exchanges and figure out their infra quirks. Don't bother with mid tier exchanges to start.
8/ There is no free lunch. Remember: pnl is always the market paying you to correct an efficiency. What are you uniquely positioned to fix? When starting out you should probably find an opportunity that is unattractive for the big boys. Something that doesn't scale with capital
9/ And finally, bias towards doing. You'll only need pointers when starting out Once something works, there will always be more ideas than you have time to explore. Each answer will lead to more questions, branching infinitely. By choosing a path you'll naturally find your style
@chameleon_jeff Hi, just curious, what language and what library you are using for websockets?
@kdcd22802757 We use rust. Tungstenite is a pretty good ws library
@chameleon_jeff @threadreaderapp unroll
@chameleon_jeff @memdotai mem it
@chameleon_jeff Awesome
@chameleon_jeff Data, can you generate 1 TB of data a week?
@chameleon_jeff @threadreaderapp unroll #HFT
@chameleon_jeff Statistical arbitrage medium Frequency Trading should do wel