Pls help me understand option market making

I'd like to understand the mechanics of options market making and how market makers manage their risks each day. It's been a matter of great curiosity for me and I think there's much to be learned by understanding this.

I understand that this has all been computerized, but let's assume we're back in the old days and humans did this.

Firstly, the job of setting the bid-ask spread at each strike. I assume the MMs use an options model of some sort to determine a fair price for the options. Anyone knows what model is used nowadays?

Since IV is a derived value, how do you determine what value to use, say, when a new product with options just gets listed?

What kind of algorithm or formula do MMs use to increase the IV (i.e price) when demand increases? Similarly, what formula is used when supply increases and IV decreases?

How do MMs determine how far to set the bid-ask?

Now about managing risk. Let's take a very simple example. I've just sold a couple of puts to a customer. I'm negative deltas so I immediately go short enough underlying to get roughly delta neutral.

Although I'm roughly delta neutral for smallish moves, a large move will still offset my deltas because of gamma. What happens if a large move happens and I'm still left holding that short put? Do I keep hedging with the underlying?

This put in my inventory also carries Vega risk. I assume they hedge it by buying options to balance this out? What if I want to buy, but no one's selling at my prices?

Maybe we can start with this, thanks.

 

You should read "Option Market Making: Trading and Risk Analysis for the Financial and Commodity Option Markets" by Baird. it should answer a lot of questions for you.

 

There are a lot of other threads that talk about this as well.

 

Thanks MTE. I have the book, but I don't remember him talking about the actual mechanics.

 

I can speak to the good old days. The bid ask spread was set by the exchange and the MMs could make it narrower if we wanted. When a new product was listed you would look at the historical vol of similar stocks, but you obviously couldn’t know for sure so you were basically guessing and then just adjust as the order flow came in. That was a key theme: order flow. People buy options from you, then raise vol, and when they sold them to you then you lower vol. It was not all that complicated, and there was no algorithm. But each pit was different, a large pit with an important lead market maker would mean most MMs would take their cue from him and other pits only had a few MMs or even just one. So pit politics would play a roll in all decisions.

Hedging your greeks was a completely personal and firm specific issue. Some MMs were comfortable taking significant risk and others would go out of their way to hedge everything constantly. If you ended up lopsided you would just change vol enough to get someone to trade the other side and it was common to lean one way or the other with whatever market was on the screen. For example I might yell out one to a quarter for a market but I might be 100 contracts bid and only sell 5. If your stock was listed elsewhere you could try to trade with the other exchange if they didn’t agree with your vol, but usually they would just move. You were also free to trade with other MMs, but it was definitely possible to end up in position you did not like and you just had to work at fixing it.

Keep in mind that as MMs back then we did have significant edge but that by no means made you a guaranteed success. There were a lot of traders who tried and failed at it. You also had to make a quite high percentage return on your haircut to cover all the costs and pay a large percentage of your profits to whomever was backing you. So just being profitable was not enough.

 

Great post opt789, thanks for sharing, those must be really interesting days.

There's something I'm not quite clear on. I think there were no such thing as a spread orders in the old days? So nowadays, when a retail customer places a spread order, let's say a put credit spread, and the customer is hoping to get filled at mid, his order isn't actually reflected on the bid and ask of the actual options themselves.

Now I assume the bid-ask spread is the minimum edge that the MM wants in order to do any trade, so unless someone sells to him at the bid and buys from him at the ask, he's not going to fill this spread is he? And since the spread order doesn't hit the books, there's really no reason for him to fill it?

So most of the time, who is the one who fills spread limit orders priced at the mid-price, the market maker or another customer taking the opposite trade?

Thanks.

 

Hlpsg,
Yes there was spread orders, all kinds back then. The order would come to the floor via a broker who would walk into the pit and ask for a market and we would respond. If it didn’t fill it would usually just sit with the resident pit broker who would hold it and try to fill it if he could at some point. You are right that spread orders are not reflected on the screen markets.

If we didn’t want to fill a spread order at the time then it would just sit there until the stock moved enough where we wanted to do it (which was the most likely scenario), another customer order came in on the other side (which was pretty rare in the average stock), or the broker could shop it around to off floor firms. It was not uncommon back then to lean on an order and just wait for a favorable price in the stock before we would fill the option order.

MMs could step in at any time and fill any order at any price. Independent traders would usually try to wait for favorable prices, but firm traders who just did what their company’s handheld told them to do would be more likely to step in and cut the markets. In my pit we wanted more volume so we would try to fill all orders that were reasonable, while the guys in the SPX pit routinely go out of their way to screw people – again these types of things were pit specific.

 

opt789,
Would you say that most options orders that came into the pit would be:

A) filled at mid-prices, or
B) buy orders filled at ASK, sell orders filled at BID?

What would be the percentage between A and B?

Also, how often do you receive market orders from customers that come into the pit?

Thanks.

 

I heard somewhere that a spread order may be filled not only against the exact opposite spread order (i.e., long put credit spread against short put credit spread with the same strikes) but also there is implied pricing functionality which allows a spread order to be filled against individual legs, so that even if there is no opposite spread order in place, if the price of the individual legs fits to the aggregate limit price of the spread order it can also be filled that way. Is it correct?

 

Hlpsg,
As I said, my knowledge is from the old days so I can’t tell you exactly what happens today. However, I can tell you when I did it that most orders would be filled close to or at the bid/ask (good for MM bad for public). Market orders were definitely less common than limit but we certainly got them. To address your question in general, you have to have a lot of experience with the specific underlying to know how to get filled on your option orders. What you do in SPX is no the same as NDX, and what you do in the Euro is not the same as oil. To trade options you have to know what fair value is at all times and have a lot of experience in that specific underlying. I see public orders everyday in things that I am watching that were put in at stupid prices because the person who put it in didn’t know what they were doing.

Dragonman,
A market maker (or anyone else who sees it) can fill your spread order at any time and for any reason they want. I occasionally get really good fills on spreads, and other times I end up giving up because I can’t get anyone to do it. When I was on the floor we would fill any order if it matched up on the screen of any other floor. So if the pit broker was holding a spread that we didn’t want to do at the time, and another single order came in for one of those legs, the pit broker could try to fill one side of the spread with the single public order and the other leg with us.