Thursday, October 15, 2009

One Observation Regarding the Distribution of Trade Sizes in ES Futures

As an initial exercise, I took a look at the first 10 minutes of trading in the S&P 500 Index (ES) futures. We transacted 82,782 contracts. Here's a breakdown of 10-minute volume by trade size:

1 lots: 7341 contracts
1-5 lots: 24,847 contracts
6-10 lots: 10,791 contracts
11-49 lots: 25,997 contracts
50+ lots: 21,147 contracts

As reported earlier, the CME is reporting volume in smaller blocks, based on the identity of the counterparty. Thus, if you sell a 50-lot, this would be reported as 50 one-lot trades if you transacted with 50 different counterparties. Conversely, it could be reported as a single trade if one counterparty took the other side.

The distribution of trade size from the little exercise above suggests that a significant portion of volume is transacted in sizes of 11+ contracts; about a quarter of volume transacted in 50+ lot units. In future exercises, I'll examine other time periods during the day.
.

8 comments:

Matt Fahmie said...

Excellent post

DaveO said...

Thanks for your trouble brett.

David.

Brian said...

If you are a price action or time and sales tape reader, doesn't this turn your experience upside down?

The counterparty is responsible for what size we see on a time and sales, before it was the initiating party.

So now, there is a big seller ripping through the market, you don't see him on the T&S UNTIL he meets a responsive buyer of equal or greater size. So in theory you think the move is gaining speed cause big sellers are on the tape, but the reality is the move might be ending because the big seller was already there and finally hit the wall of a bigger buyer?

This doesn't particularly apply to my style but am I missing something?

OKL said...

Doc, should the second line read "2-5 lots" instead of "1-5 lots"?

DaveO said...

I don't think you are missing anything :-)

bullballs02 said...

Appreciate you digging into this.

otterby said...

a plethora of thanks..

steveplace said...

i want to try this study:

make 2 separate vwaps, the first solely for large block trades and the other for small block trades. see if there's any differential, and see if there's any signal to be had

any thoughts?