Option Spread Execution Slippage

[optionshedge]

Hi,

I frequently trade multi-leg option spreads (i.e. condors / butterflies) on indexes, such as NDX, RUT, SPX.

Occassionally, I am able to get better than mid prices on my option spreads. But more often than not, many of my option spread orders don't get executed, even when I give up substanial edge from mid-prices (especially on the NDX).

For instance, I would submit concurrent butterfly orders on the NDX using different strikes and ratio's. One spread order gets executed at better than mid-prices, while the other order doesn't execute even after giving up $0.30 edge from mid.

What are the factors that determine why one spread gets better than mid-price execution versus no execution for the other spreads? Are there any rules of thumbs to determine which spreads are likely to get filled versus not filled?

For instance, the legs of one spread order may have lower bid/ask spreads (higher volume, higher OI, closer ATM, etc..) than the other, and hence may get filled at better than mid prices?

 

[1245]

One thing to keep in mind is that the mid point is meaning less to an option market maker that runs their own skew and values. They look for edge based on thier values not the mid-point. They majority of your executions will likely be with one of those electric MM. They have the ability to monitor the COB for edge. Very few options traders watch the book with their eyes, then trade with you. You might get lucky from time to time and trade with another customer.

EG Let say you are looking at a spread where the two markets are:

10.00 x 9.00 (mid-10.50)
5.00 x 6.00 (mid- 5.50)
Based on the mid points you will think the spread is worth $5.00. Let's say that is correct. Now a bid comes in on the second leg and the market is now: 5.25 x 6.00 with a mid point of 5.625. Now the spread mid point is 4.875. The spread is still worth $5.00 to them. If you bid $5, above the mid point, they won't sell it, because there is no edge to them. This is why you have to be careful making assumptions about mid point values in options that are very wide without customer flow.

1245

 

[surfer25]

Quote from 1245:

One thing to keep in mind is that the mid point is meaning less to an option market maker that runs their own skew and values. They look for edge based on thier values not the mid-point. They majority of your executions will likely be with one of those electric MM. They have the ability to monitor the COB for edge. Very few options traders watch the book with their eyes, then trade with you. You might get lucky from time to time and trade with another customer.

EG Let say you are looking at a spread where the two markets are:

10.00 x 9.00 (mid-10.50)
5.00 x 6.00 (mid- 5.50)
Based on the mid points you will think the spread is worth $5.00. Let's say that is correct. Now a bid comes in on the second leg and the market is now: 5.25 x 6.00 with a mid point of 5.625. Now the spread mid point is 4.875. The spread is still worth $5.00 to them. If you bid $5, above the mid point, they won't sell it, because there is no edge to them. This is why you have to be careful making assumptions about mid point values in options that are very wide without customer flow.

1245

Hi 1245,
Are you saying that unless the midpoint is determined by the MM, you can't consider it a true midpoint? i.e. if a non MM order changes the midpoint, the apparent midpoint is just an illusion. Also, what percentage away from midpoint generally becomes attractive the the MMs?
Thanks.

 

[optionshedge]

Since a spread has to execute all its legs on one exchange, which is a better reflection of the fair value of the spread:

1) The midpoint of the spread using NBBO prices, where different legs prices may be from different exchanges.

2) The midpoint of the spread using prices for all legs on just one exchange.

3) For example, if routing the spread to ISE COB, would it be better to use the ISE prices only to calculate midpoints versus NBBO prices? Or CBOE prices if routing to CBOE COB?

 

[1245]

Quote from surfer25:

Hi 1245,
Are you saying that unless the midpoint is determined by the MM, you can't consider it a true midpoint? i.e. if a non MM order changes the midpoint, the apparent midpoint is just an illusion. Also, what percentage away from midpoint generally becomes attractive the the MMs?
Thanks.

The mid point of a spread or single leg is simple the bid/ask average of the NBBO. In options with tight spreads like near term SPY, it typically represents a consensus of the current market as to fair value at that moment in time. In options with very wide spreads, adding a better bid or offer, only tightens up one side. It does not make that option worth more or less because a bid or offer, that no one will trade with at that moment, enters the market. You have to watch the current Ivol for that mid point and make sure it fit in your parameters.

 

[newwurldmn]

What 1245 is saying is that the market making firms have their own fair values for the options. It often isn't the mid point (especially with non-ATM options). They will only trade with you if they have edge to that fair value.

So you will get executions based on where your price is to their fair values, not the listed market prices.

This is especially true in the index markets where listed spreads are comparatively wide and the market maker valuations are incredibly precise.

2.00 at 3.00 might have a fair value of 2.2 to 2.8 for a particular market maker. If it's 2.25, they will often pay 2.2 and sell at 2.3 but you won't know that until your order enters the market.

 

[optionshedge]

Is there any way for us retail traders to determine what market markers are pricing as their fair values for option spreads, maybe by looking at certain factors (i.e. order flows, volume, OI, etc..)?

Or is this a un-predictable / random process?

 

[Brighton]

Quote from optionshedge:

Is there any way for us retail traders to determine what market markers are pricing as their fair values for option spreads, maybe by looking at certain factors (i.e. order flows, volume, OI, etc..)?

Or is this a un-predictable / random process?

I don't have a direct answer to your question, but it's come up before and someone mentioned that if you're trading size and in a limited number of names (so you get familiar with them) you can come up with your own volatility model, and throw your hat in the ring as a market maker of sorts. Decent options analysis software will let you manipulate the curve, change the underlying's price, change the ATM IV or individual IVs and so on. Your results might be different than the quoted mid of the bid/ask - et voilà! - you now have edge!

 

[optionatrix]

I trade the SPX. I don't start with the mid. I always look at the (near) equivalent SPY position and closeby positions first, and compare the IVs. There are differences in terms of price, dividend, and exercise style, but the SPY mids gives me a general idea of the price I should put on for the SPX spreads. Sometimes I end up putting limit orders better than the shown mid based on what the SPY is showing.

My SPX trades usually get filled. Sometimes very quickly and other times the SPX price has to sway a bit, but not a lot.

 

[optionshedge]

Thanks for your reply.

But the question is less about how to define your own edge, but more about whether it is possible to systematically / algorithmically estimate the prices at which a given spread can be executed by looking at various factors?

Given a series of spread candidates, how can one forecast which spreads are likely to execute at better than mid versus worse than mid? Or is it all just random and unpredictable?

Quote from Brighton:

I don't have a direct answer to your question, but it's come up before and someone mentioned that if you're trading size and in a limited number of names (so you get familiar with them) you can come up with your own volatility model, and throw your hat in the ring as a market maker of sorts. Decent options analysis software will let you manipulate the curve, change the underlying's price, change the ATM IV or individual IVs and so on. Your results might be different than the quoted mid of the bid/ask - et voilà! - you now have edge!