Mark Aron Szulyovszky @almostintuitive

Building a unique product in the quantitative finance space (in stealth mode). almostintuitive.com Berlin, Germany Joined January 2012

Tweets

730
Followers

493
Following

508
Likes

121

Ruizhe Jia @RuizheJia

3 weeks ago

Binance is arguably the world's largest crypto exchange. Feb 2026: Polymarket lists a small 5-minute Bitcoin contract. Since then, Binance order flow spikes in the final seconds before settlement; prices revert soon after. New paper with David Dai and Shihao Yu ( @ShihaoY ): Settlement Manipulation in Prediction Markets (papers.ssrn.com/sol3/papers.cf…)🧵 The finding, up front: Traders push Bitcoin's price in the final seconds to decide the contract. That push makes the price less informative, yet Bitcoin's market more liquid. Market makers are largely insulated; ordinary traders lose $7.6M in two months.

73 124 1K 250K 926

View Details

Daniel Szemerey @szemyd

2 weeks ago

It can be lonely being a quant; ie especially if you are behind large walls that forbid sharing info (for good reasons though). I think the biggest added benefit of running a service like aperiodic or aperiodic factors is that we get to talk to a lot of teams on a daily basis

0 1 1 37 0

View Details

Daniel Szemerey @szemyd

2 weeks ago

Buying makes more buying likely: not for seconds, but for days. It seems the tape doesn't actually forget the morning, it simply fades it out slowly.

1 1 2 151 1

View Details

Mark Aron Szulyovszky @almostintuitive

3 weeks ago

Putting together the content for Aperiodic has been a nice excuse to read and re-read some of the most interesting market microstructure papers. I also wanted to write about the (great) variance in the quality of these papers, but I'll save that for another day!

0 0 1 20 0

View Details

Mark Aron Szulyovszky @almostintuitive

3 weeks ago

We also serve something we call Spread/depth ratio: low values means genuinely cheap liquidity. High ratio means a narrow quote is only the surface, and it will not survive contact with real flow.

1 0 1 25 0

View Details

Mark Aron Szulyovszky @almostintuitive

3 weeks ago

I thought I had a good intuition what "spread" is, but digging through the literature turned out to be surprisingly enlightening.

1 0 0 82 2

View Details

Mark Aron Szulyovszky @almostintuitive

4 weeks ago

why did I choose handles and company names I could never properly spell or type without making a mistake? it's only muscle memory that saves me, I could not deliberately type out any of them. probably dyslexia, right?

0 0 0 52 0

View Details

Mark Aron Szulyovszky @almostintuitive

a month ago

Try basis, order book imbalance, flow metrics. We cut headaches by 90%, and found reliable alpha that we can build upon, without worrying about whether any of it is actually traceable.

0 0 0 27 0

View Details

Mark Aron Szulyovszky @almostintuitive

a month ago

In the end, the operational overhead could never justify the alpha we could find: not saying on-chain metrics are not fun, but derivatives in crypto are 80% of the volume.

1 0 1 30 0

View Details

Mark Aron Szulyovszky @almostintuitive

a month ago

Why did we end up specializing in market microstructure / liquidity / flow metrics? When we originally started exploring crypto, we tried every data source imaginable: on-chain, social sentiment, etc. We got pretty deep, and had working pipelines to ingest data from...

1 0 2 50 0

View Details

Mark Aron Szulyovszky @almostintuitive

a month ago

I didn't think 6 months ago that I could spend a weekend on an LLM-assisted deep dive into the message brokers' internals we use (NATS Jetstream) and speed up the hot path (specific to our usecase) by 2x. Interesting times!

0 0 0 34 0

View Details

Mark Aron Szulyovszky @almostintuitive

2 months ago

It'd essentially moving the bidding process on-chain. Why? Probably every setup is better than the current monopoly, right?

0 0 0 30 0

View Details

Mark Aron Szulyovszky @almostintuitive

2 months ago

where the priority of the domains are set by token holders. Search is burning a very small amount of the token. Holders are incentivized to be both promoting their own agenda but keeping the usability of the engine. Competitors will enter and bid for higher ranking

1 0 1 45 0

View Details

Mark Aron Szulyovszky @almostintuitive

2 months ago

Interesting that there are a couple of obvious-sounding "crypto projects" that were never seriously tried, let's take "decentralized search engine" as an example: Naively, imagine a blockchain-based keyword + vector database and query engine...