The reason I posted this is that I got a few emails questioning my 80% program trading and 3 % retail trading comments yesterday. Since the frequency and volume of program trading have shown a pattern of increasing over the past 10 years I'd be surprised if this year's numbers aren't even higher.
This data is a rehash of a series of studies I ran last year (starting Sept, 1 2007) exploring the effects of order flow masking algorithms, iceberg orders, and backchannel order executions. Interestingly enough, retail orders typically print de facto, whereas pro traders' orders are shown in increments of 100, 200 and 400 shares. Watching level 2 time and sales stream by, those 1000 share prints are probably retail orders, not the big dogs, so it's important to keep track of the total volume per bar rather that the T&S when trying to assess momentum.