Not sure what to make of it – there were a $hit load of 100 share trades on the THRD and NMS exchanges from 14:30 through after hours.
The low trades on T&S were around $69 and in 100 share blocks. It started fading from $84 at around 14:30 and started lifting of $70 around 14:46.
If there was a huge trade that came as a limit order below the bid, the market maker doesn’t have to show it on T&S – only that a trade occurred at that level. It lifted from ~$72 around 14:46 back to ~$83 by 14:58 and floated into the close settling around $84.
Here’s where it gets interesting – At 16:48:51, the THRD lights up at $73.25 and there are literally pages upon pages of 100 block trades going through – all at $73.25. There are close to 2000 trades executed in 6 seconds. Then at 16:48:57 the THRD jumps to $84.24 and about 300 trades go through. After that, most of the trades originate from the NMS exchange and are the standard Best Bid/Best Ask sell and buy drill – again, all around $83-$84.
Rumors are flying that it was a P&G trade that started the landslide. Their T&S looks just like 3M, except the low trade came in on the NMS exchange at ~$39 at the same time the THRD was executing at $60 – all right around 14:47:15. By 14:48:11 it had recovered to ~$60.
I’d like to know who pocketed the $11 spread on the 3M trades from 16:48:51 to 16:48:57.
It still looks like classic programmed sell/buy response to an external price stressor to me. In other words, when all is said and done, the market was down 350 today – NOT 1000.
It will be interesting to see if any of these trades get whistled out overnight.
I support your dig through the data!
However, I might note a couple of things here…
- The market was sliding all day but really only picked up steam (with respect to price and volume) for three 5-minute bars starting with the 1:35 pm bar which is well before some of your T&S data.
- There was some accelleration in volume for the last two red bars but the really big volume was for the nipple bottom candle and the next greenie during the ride back up. The volume wasn’t on the downside, but the upside.
- Where you note 2,000 trades of 100 blocks as representing a lot I might note that this is only 200,000 shares which, in the scheme of things, is not enormous volume.
What I take from this is that for some reason the bids got pulled and it wasn’t super high volume killing the market. Next I’ll note that the enormous volume showing up during the instant bounce represents far more than a simple unwinding of the prior trades. A huge amount of liquidity came roaring in.
Where you seem to have already determined that this only represented the normal workings of a free and open market system I am not yet convinced that we have enough forensic evidence to make any determinations about what happened or where all that market support came from.
I remain open to the possibility that this truly represents the normal workings of finely tuned algorithms playing the circuit breaker angle but also something else like an official rescue operation. I simply don’t have the evidence to really say anything one way or the other.
Except I can say that yesterday was highly unusual and I do have enough evidence to call B.S. on many of the so-called explanations circulating out there like the idea that traders at Citi use text, not numerals, to enter trades.
More data! We need more data!!
Morning Chris –
Even looking at Volume At Price (VAP) charts only further muddies the water. The vast majority of 3M’s VAP rolled up by the time the price dropped from $86 to $82. There was very little relative VAP during the fall to $69.
Pretty much the same chart for PG except one of their VAP lines runs all the way across the chart!
Absent any other explanation, you could conclude that some consortium of schmucks dumped a load of PG and 3M forcing the drop so they could bottom feed off the dip. Except there is no volume at the low trade price.
If I could figure out how to hang a picture of the PG and 3M VAP 1 minute chart I would – even that is cumbersome since most of the volume of 3M’s ride happened over a span of about 6 seconds.
I’ll keep digging
It looks like Erik’s explanation is as good as anything else floating around right now. And I am quite certain I don’t want to examine any of that data.