Author: Dave
Why do I always have bad luck shortly after buying altcoins? Why couldn't the seemingly invincible market makers handle the sell orders on 10/11?
Why did every market maker I talked to after 10/11 casually say they didn't lose much / even made money that day? This article will introduce you to the market maker's order book and order flow.
1. Limit Order Book (LOB)
The market maker's battlefield is not the candlestick chart, but the LOB.
Core concepts:
-
Depth: The volume of orders at each price level.
-
Tick Size: The minimum unit of price movement. In a high-frequency environment, tick size is crucial for queuing strategies.
-
Price Improvement: When you provide a price better than the current National Best Bid and Offer (NBBO), you create value for the market.
For example, if the current BTC bid is 100k and the ask is 110k, if you offer a price of 101k, you narrow the spread and create value for the market.
While writing this article, I was thinking about how to explain this section to you all. Later, I felt that nothing would be clearer than showing a real order book. Let me do a little promotion for a friend at BG. Using BGB as an example:
The above is the BGB order book at the minimum tick size. We see the minimum is 0.001. The current depth is only a little over $1000, the spread is controlled to be very small, hitting the minimum tick size. Simultaneously, the depth is distributed in a "Christmas tree" shape—the further from the spread, the larger the order size. However, when we increase the tick size a bit, we find that the depth should be distributed in a horn shape: the closer to the spread, the larger the order size and the more liquidity; the further away, the fewer the orders (one of the reasons for the liquidity vacuum on 10/11).
Speaking of Christmas trees, this article's scheduled release time is 12.24, Christmas Eve. Merry Christmas to everyone!
2. Market Maker Profit Source: Spread
The spread is not just profit; it is primarily composed of three costs:
-
Order Processing Cost: Exchange fees, hardware latency, manpower.
-
Inventory Risk: The risk of price moving against you while holding a position, which we covered in the previous episode.
-
Adverse Selection Cost: This is the core—when you trade, the counterparty might have information you don't. This means you might get tricked by an insider.
-
There are also three types of spreads: Quoted Spread, Effective Spread, and Realized Spread. The Quoted Spread is the easiest to understand: the ask-bid difference. The Realized Spread measures the actual profit the market maker keeps after price adjustments: 2 x (P_trade – P_futuremid), which includes the future mid-price, somewhat like considering opportunity cost.
3. Order Flow
Order flow is the stream of orders received by the market maker. This is a deep topic. Market makers perform various operations on order flow, such as hedging, matching, adjusting quotes, etc., in order to manage their book (ledger?). This involves many professional concepts and operational techniques, and can even involve legal issues, such as agency trades not being able to be counterparties to principal trades due to conflict of interest. This article will only introduce the concepts of Order Flow Toxicity and VPIN. If any MM boss hires me, I'll update my followers on professional order flow management.
Toxic Flow refers to orders from informed traders who know the price is about to move, thus causing realized spread losses because they know P_futuremid through insider information. So, as market makers, we also need to be careful not to be tricked by toxic insiders.
Noise/Retail Flow comes from retail investors or funds passively adjusting their weights. This is the market maker's favorite "food".
To protect themselves, market makers' anti-toxicity mechanisms adjust quotes. A simple way to prevent "toxicity" is to assume all aggressive orders are toxic. Assuming a buy order, the MM immediately lowers the reservation price, shifting the entire quote down. This also answers the question left in the previous chapter: why do we always buy at the high? Because market makers adjust their quotes for risk control. But the smart you will surely ask another question: what if an insider buys a large amount? What if they rely on their information advantage and stubbornly impact the spread head-on? Yes, this might be what happened on 10/11, and it's also the reason why a large MM with billions in assets couldn't hold the line.
Core metric: VPIN (Volume-weighted Probability of Informed Trading).
VPIN ≈ The probability that the market maker is currently being "continuously hit by one-sided flow" in the market. When significant one-sided pressure appears, MM inventory accumulates one-sidedly, and the mean reversion assumption fails. At this time, the MM will pull quotes (withdraw orders), temporarily stopping providing liquidity to wait for the order flow to return to symmetry. But what if the order flow doesn't return to symmetry? Or if the order flow deviation is too extreme and directly causes a blow-up? This is the 10/11 disaster. I'm thinking about doing an episode on how exchanges made money on 10/11. We'll see.
Back on topic, after a VPIN anomaly, MMs will pull quotes (withdraw orders) or widen the spread (to earn more service fees to compensate for price losses), and also reduce size to control the speed of inventory accumulation.
This episode is the end of the first part of our market maker story. From a retail perspective, the truth behind the legendary market manipulation by "whales" has been revealed. Next, I will introduce some more "institutional" topics from the MM's perspective. Hold on tight.
Anime episodes often end with a preview for the next episode, so let's do one for this article: If we entered the world of Jujutsu Kaisen, and order flow is "cursed energy," and quote operations are "cursed techniques," then in the next part, we will look at the "Domain Expansion" in market making.
To find out what happens next, stay tuned for the next episode.
