Why wait for the slow train?

Here’s an insight about the physical limits of MEV distilled into a short parable. The moral of the story is that to get the most benefit from acting on combined information, it may be necessary to commit not to acting on partial information, even when the latter is available more timely. There’s a fundamental exclusivity here: the authority to act on the information can’t reside in both places at once.


In a bustling industrial town, there were two factories, Factory A and Factory B, each producing a crucial component for a popular gadget. The town also had a central market where traders would gather to buy and sell shares in these factories based on their production output.

Two rival trading firms, Firm A and Firm B, each had a pair of traders: a junior trader known for their speed, and a senior trader renowned for their wisdom and negotiation skills.

Every day, as the factory whistles blew to signal the end of the production shift, the junior traders would dash to the fastest train, the Green Line, racing towards the market. They carried with them the day’s production figures from their respective factories.

The senior traders, however, would leisurely board the Purple Line, a slower train that took a scenic route to the market. Interestingly, this train had a special carriage where the senior traders from both firms would meet.

Now, the junior traders, despite their speed and eagerness, were bound by a peculiar company policy. Upon reaching the market, they were not allowed to trade immediately. Instead, they had to wait for their senior colleagues to arrive on the slower train.

This policy often frustrated the junior traders. “Why must we wait?” they would grumble. “We have valuable information that could make us a fortune if we act now!”

Little did they know, something important was happening on that slow train.

In the special carriage, the senior traders from Firm A and Firm B would engage in intense negotiations. They knew that while their individual factory information was valuable, combining it could reveal insights far more precious.

However, these negotiations were delicate. Neither senior trader wanted to reveal their information without assurance of how it would be used. They needed time to build trust, to bargain, and to strategize.

One day, a wise old trader explained the situation to the impatient juniors:

“Imagine Factory A had a great day, but Factory B struggled. If you both traded on this partial information, Firm A would buy, and Firm B would sell. But what if the combined information showed that the struggles at Factory B would soon affect Factory A? By waiting for your seniors to negotiate and combine their knowledge, you might both decide to sell, avoiding a costly mistake.”

He continued, “The senior traders only engage in these valuable negotiations because they know you won’t act on your partial information. If they thought you might trade early, they’d rush to the market too, and we’d lose the chance to make better-informed decisions.”

The junior traders began to understand. They realized that their waiting allowed for a deeper, more strategic approach to trading. By holding off on immediate action, they were enabling a process that could yield far greater rewards.

From that day on, the junior traders waited patiently at the market, knowing that the slow train was bringing not just their senior colleagues, but a wealth of carefully negotiated, combined wisdom that would lead to more profitable trades for both firms.

9 Likes

OBJECTION :military_helmet:

I don’t buy the argument. You are making the assumption that one can both move through space and do computation at the same time. Perhaps this is possible within the bounds of physics, but not with the tech we have today.

Since things happen sequentially (you can’t talk on the train), the fastest way to get the trades based on shared information on the market would be to take the separate fast trains to the market and then to negotiate there.

:star2: A fix :star2:

You can get the same point across by having two destinations (e.g. two trading venues). Then the optimal point would be smack bang in the middle (assuming we value the sources and markets equally).

3 Likes

This is a great point, computation and transmission have to occur in sequentially, you can’t really compute on data while its in transmission. This is an essential breakdown between the real world analogy and the computing scenario we care about.

I like the suggestion there may be other simpler explanations that have a similar outcome, like if there are two markets then it’s helpful to merge information at a location between them… Not clear the compute latency of the merging itself is being relevant in that case though. So for now Im considering the specific idea in the op refuted.

image

3 Likes

But, but… I liked the story about waiting, maybe for a different reason than the intended moral. So I’m inviting this thread to go in a different direction.

I want to live in a slow economy. Why is it bad to wait 10 minutes for a transaction to be confirmed? Faster transactions are driving us into a race, that just benefits faster and faster speculation. This is doomed. What if we design a system that moves at a different pace? Will we have time to be more wise, like in the story?

Let’s say that a transaction has to collect attestations from 3 different continents and from 3 nodes running different client software. More like onion routing than amazon colocation. Who would benefit from a shift like this? What would traders do?

The follow up question I want to focus on is this (from discuss w @fiiiu and @Quintus ):

  • is there another simple model, similar to the physical train model above, but that doesn’t rely on computing while travelling, that ends up with a similar conclusion, that even when colocation is possible it may tend towards diversity rather than winner take all?
3 Likes

You mean “diversity of location” right?

The cooperation of traders is baked into your description of the market afaict.