This must be old news to guys who do microstructure bid-ask work in ES. I looked at such data for the first time today in any granularity. It seems there is no relationship between trade size and inside bid-ask size changes. Caveat - this is IB snapshot datafeed so not accurate, but still bid-ask size doesn't change even when large orders execute. So, I guess it means a lot of these must be icerberg orders ?? Wondering!! It also means that if inside ask is X cars during the morning hour, it is certainly possible to probably buy 2 to 4X the displayed size without moving the market. Just wondering - if someone experienced decides to share some details on market microstructure mechanics of ES, that will be great! I trade using 5 min charts with holding period in hours, so my market microstructure observations are just an academic curiosity at this time.
Why would you even want to try drawing conclusions from one day of observation? As an analogy, it was about 50 degrees in NYC today. Should I therefore conclude the temperature in NYC every day is about 50 degrees?
I wouldn't bother doing such an "academic analysis" until getting better suited data. IB is not suitable for such and a better data source is going to cost you a lot more money per month.
We are looking at high frequency data. One day of data is not enough to make a strategy, but it is definitely enough to notice important facts about the market. Just putting forward here what I saw. Hoping someone will be willing to post a bit more - what I have posted would be the work that professional microstructure guys would do in their first afternoon on the job. So, I know its pretty damn basic - but I saw it for the first time and found it interesting - so posted.
No response! I guess I should have posted this observation on more quant kind of sites like nuclearphynance or wilmott.
Observed same for NQ. Inside market is not driven by market orders but by bid/ask being lifted. So IMO watching inside market or DOM is useless. And it hurts your eyes and head. I will generalize that asking ANY intelligent question here is useless.
OK, so to amend my analogy, I could take the temperature of NYC every millisecond and average them out to find that it was 50 degrees. Should I therefore conclude that it is 50 degrees every day because I have hundreds of thousands of temperature observations? I admit that I don't find microstructure really interesting at all, so maybe this is just the wrong thread for me to start. It seems to me to be a place where people who can't find an workable strategy go to basically look for a needle in a haystack. And forget about making something profitable at the retail level, where commissions and slippage will kill you.
Anyways, I am hypothesizing that this phenomena we are seeing could mean few things: 1) The size that can be transacted at a level in ES cannot be computed by looking at order book. Rather the actual volume that is getting done will probably offer a better yardstick 2) Since ES is closely linked to stocks, no arbitrage condition means that dollar value of the volume transacted per minute in 500 stocks is probably more reflective of the dollar value of ES that can be transacted per minute rather than ES order book (since obviously a lot of iceberg orders are present there) 3) Above 2 points also mean that one of the logic behind pulling in and leaving bid-offer by participants would be driven by keeping track of what is happening in those 500 stocks. 4) I would venture to say that during some times of the day, ES leads stocks (especially when we are approaching big support resistances on ES) - whereas on more fundamental news driven kind of days, it is stocks that drive ES. No evidence, just hypothesizing.
Only thing I can add is that options and futures on options traders bring to the party a lot of hedging volume. Also index funds must buy/sell futures to balance out the positions they will have to take after the close to cover customer redemptions/purchases, which they will unwind the next day when it is most advantageous to them. I will also add the unpopular position that volume doesn't mean shit, price will be lead around by the nose like a prized heifer on invisible volume as it suits the suits.