Sound the Retreat

Continuing from the previous post about market makers, let's look at an example.

Let's set the scenario. Let's say this mystery contract ticks in nickels (0.05) and I always want an edge of 0.12. Also, I am willing to do 10 lots (contracts) at a time. Finally, our algorithms tell us that the contract is worth 6.00. So, what is our bid and what is our offer?

Since we want at least 0.12 edge, the best bid we can put in is 5.85 bid (since 6.00 - 0.12 = 5.88, but it ticks in 0.05, so the highest I can go is 5.85). Likewise, we want 0.12 edge on our offer as well, so our offer would be 6.15 (again, 6.00 + 0.12= 6.12, but it ticks in 0.05). So, we are 5.85 at 6.15, 10 up. And to make it easier, we can show:

1

Great. So, we show our market in the exchange and we wait.

A buyer comes into the market and is looking to buy 7 lots. We show him our offer and he decides he can pay 6.15. Since we are only comfortable doing 10 lots at a time and we already did 7 on the offer, we now have 3 left on the offer. So, our market (bid-ask spread) looks like the following:

2

Let's continue. Another buyer comes in and buys the remaining 3 lots we have sitting on our offer. That means, we have no more lots to offer at 6.15. Now, we "retreat."

What is retreating? Retreating[1] is the consequence of updating the perceived value (theoretical price) of the instrument.. In doing so, you will set up a new market (bid-ask spread) that will biased into flattening out your position. For instance, the more you buy a contract, the less it becomes worth to you (and likewise, the more you sell something, the more it will become worth to you). In treating the value as such, you will become more and more aggressive to exit your position. Let us revisit the above scenario to drive the point home.

Let's say we retreat 3 cents. Like I said before, since we sold, we increase the value of the instrument. Then the new theoretical price will read 6.03. Since I am always seeking at least 12 cents in edge, my bid will now be 5.90 (6.03 - 0.12 = 5.91). Further, my offer will be 60.15 (6.03 + 0.12 = 6.15). So my market would look like the following:

3

There are a couple of things to notice. First, notice that we didn't retreat our offer. We are still at 6.15 (still have an offer at 6.15). But, notice our bid, its moved up from 5.85 to 5.90. We are more aggressive to buy back our short position. And the more aggressive we get, the more likely our bid will be filled.

Something else to note is that the amount we retreat will reflect our appetite for risk. In the previous example, we simply retreated 3 cents. That means, we would have to retreat 4 times (12 cents of edge / 3 cents to retreat = 4) for our initial trade to be an even trade (no profit). Let's take another look at the example but set a different retreat amount.

This time, let's say we retreat 5 cents. Using the same logic as before (theoretical value goes up 5 cents and we recalculate our market), we get the following:

4

What if we retreated 8 cents? We would get the following:

5

Keep in mind that all the retreat examples we looked at were smaller than the edge we ask for. Imagine if we retreated 50 cents (when we only asked for 12 cents). We would be doing trades left and right, but we would be giving away free money! We would be buying it expensive and selling it at a steep discount.

Retreating is an essential part of market making. In doing so, you are more likely to get out of your current position and less likely to expand on it. In our example, we sold 10 lots, and as we continue to retreat our bids got more aggressive to buy back our sold contracts and our offers became less competitive as we backed away from our initial price. Again, the amount we retreated determined how aggressive we were trying to get out of our position.


  1. The actual retreating can be done in (mainly) two different ways. First, you can change your theoretical price to reflect what the new value is. Second, you can change the amount of edge you ask for. I personally believe the former is the way to go, as the latter method proves to be not only cumbersome but deceptive of what you truly believe the value of the contract is. ↩︎