Another cool flow idiosyncrasy that may interest those flow nerds out there. Cryptocurrency markets have more momentum in the short term, which is usually dominated by mean-reversion effects. This is basically all driven by over-levered degens and algorithms that...
1/idk lol
get access to these 100x leverage futures or 20x levered futures, and unlike most would think it actually gets used. Especially if you are some sort of HFT algorithm and now you can lever up as much as you want in a super volatile market. Plus with the sorts of...
2/n
profits some of these guys make, who wouldn't lever this thing up to capacity and get in while it lasts. I will probably still attribute most of the momentum still to retail guys getting liquidity cascaded out of positions. i.e. takes loss, gets out, price moves down...
3/n
and the cycle repeats for another investor. Then suddenly we get some mean reversion algorithms or just finding a general floor and the market settles, but that's a large part of why markets for crypto are so volatile. The idea that it is because there is nothing holding...
4/n
it down to the real world in terms of value is BS and everyone knows it. Stocks can do whatever they want, the volatility complex has shown it where "hey you want low index vol bc institutional selling, but still crazy vol idiosyncratic bc liquidity is fucked, yeah sure...
5/n
have it" and that's just dispersion. (not that institutional selling is the only reason - lots of it is MM, or that liquidity is all of it, shift to passive is a big one too), but the market clearly just does what it wants. We saw that with AMC as well. They just issued...
6/n
new shares and got hundreds of Ms richer and now they aren't bankrupt and instead the C-suite is loaded. Stocks can entirely be away from reality if the market will let them, and so can cryptocurrencies. When leverage decreases and institutional buying comes into place...
7/n
then we will see cryptocurrencies settle down in terms of volatility. It has nothing to do with their lack of fundamentals. They have plenty of fundamentals, just look at blockchain data/ wallet data/ coin velocity... etc there are like 30 different figures, much like...
8/n
earnings for a company. There are growth prospects as well, such as Solana blockchain / underlying tech etc. Anyways that's my second thread on niche flow shit for the day, and hopefully you learnt some cool stuff about crypto momentum existing short term, and flow > fundamental
yes, I said there are fundamentals, yes I still believe it is multi-level marketing for nerds. No, I do not HODL, only HFT in case you are wondering.
• • •
Missing some Tweet in this thread? You can try to
force a refresh
For those of you who aren’t aware of the types of funds. When someone says CTA, they’re usually a momentum fund or capture some well known futures effects. A lot of these guys are a strange form of levered beta and not really alpha because a lot of their methods work best…
1/n
when lots of others use them so strangely enough not the most advanced or accurate model is the best model because really it’s just this greater fool theory flow. You can pretty easily frontrun this shit, because they typically ease into positions because they have…
2/n
Quite large books, and they compound their positions based on the simple rule of add to winners and cut losers (basically the entire premise behind momentum where losing money is a get out signal and making money is the get in signal). Anyways, if you know there is…
3/n
Large quantity of unique features
Really good dimensionality reduction
Ensemble everywhere!
A word on each...
When it comes to modeling everyone always goes to their favorite NNs like LSTMs etc or LGBMs and those are great, but everyone has them, and frankly, they aren't that hard to implement! Just look at Kaggle if you want an example of DS students using them everywhere...
For real alpha, you need to focus on the three most ignored areas (there is a fourth, speed, but that's not really modeling, and a fifth which I'm not telling you because I like my alpha unleaked). That sounded super guru-like, but I promise these work and I use them.
For those wondering if they are, I'll give a few comments:
T-SNE is always something I want to apply, but can never quite figure out the right way to do it. There is certainly some benefits to be had from having a basic understanding of what this all means so you can get a better chance at visualizing your features in 2D...
Much like stochastic methods, as much as I would never make them the center of a model, there is always use as a feature or ensemble. Ensemble is truly the free lunch of alpha...
A key concept for MMs is how you manage inventory. Avellaneda and Stoikov is basically the model everyone uses for this. Then there comes the offset, basically how wide your spreads are. That's your basic model of liquidity provision...
From there we get to have some fun! If you can create multiple forecasts for different timeframes (and at a super-advanced level compute speeds) you can make spreads asymmetric and intentionally hold inventory...
Entirely unprompted here, but please check out @FadingRallies. Also @choffstein's Liquidity Cascades paper (link below). The flow between MMs, passive funds, ELS, and generally the effects of reflexive dealer hedging are key to understanding this regime!
Even if you aren't a trader (I certainly am not, although I try to keep up with it all) it is still super important to understand the regime and how it all fits in from a risk perspective. You CANNOT just take the models as your risk! Eigenportfolios decay, I would know I work
with them all the time so that isn't even the perfect metric (although I do love them). Statistical models will capture some risk but at the end of the day, you choose the parameters and the distribution you feed in is key. Knowing fat tails exist is incredibly important for this
Tweeting a question I was asked/ response regarding MM:
(me adding bonus resources):
A great example of C++ HFT MM algorithms. An improvement idea I have suggested to the author but can also be attempted by interested algotraders is that a fast model like XGBOOST (there is a C++
library) is used alongside some alphas to make spreads asymmetric before traders can trade against you and you get negative edge in those trades. A large part of market making is cheaply executing alphas by trying to get inventory on the side of your predictions and also by
getting out the way of adverse conditions by making your spreads asymmetrically wide (traders with alpha against you). github.com/hello2all/gamm…