This is worth watching:

https://github.com/ethereum/go-ethereum/issues/21350

The rules dictate the game. How clients order same-price txs will determine how self-interested actors interact with the network in the future.

Currently, the rule is to choose uniformly at random.

This dictates that backrunners must submit several txs if they want a better chance of having their tx mined immediately after their target tx.

Choosing via FIFO would dictate that backrunners must compete on latency -- and could result in some very interesting network-level shenanigans, such as attempts to identify pool nodes and connect to them directly (perhaps even paying for direct access a la Flash Boys 1.0)

Choosing via "lowest tx hash" would benefit backrunning whales, as they could simultaneously sign the same tx with 1000's of different addresses (each holding enough ETH to pay the gas), and then submitting only the lowest-hash tx.

None of these is inherently bad.

But every choice of rules has a corresponding set of "optimal behaviors" that emerge once they are implemented.

So it's important to choose wisely.

When faced with these kinds of situations, I recommend stepping back and thinking first about what you want the end behavior to look like.

And once you know that, *then* look for rules that dictate those behaviors.

Having clarity on the desired behavior is important.