Trade Signals: Institutional Activity?

[GarethJ]

I have had some good success going long when I stumbled upon apparent increases in institutional buying. i.e. logic dictates that significant increases in Ave Trade Size equals more institutional activity? EXX today has a Trade Size Ave (so far) of 1018 - vs 751 and 437 the last two days. I have done a quick Google search and can't find any such indicators. OBV is useful for measuring Volume momentum, but still doesn't place a premium on large trade sizes. Does anyone use such measures and/or there any such measures "out of the box" on OST? I (naively) believe that I would like to buy when the funds are buying.

[john_1]

this is an interesting concept.. I like it.. do some research prehaps you are on to somthing..

[#3]

The problem here is there is no way to identify if the ave volume/trade size as indicated increasing in the last three days is for buying of shares in volume or selling shares in volume. If we had more accurate data showing sales vs purchasers then we could work it out. As the share ened nearly flat it is even more difficult to understand where the pressure is in trade volumes. And I know someone is going to say for every share sold there is a share bought but in what direction is the ave trade volume heading? Maybe I am not making sense....

[#4]

I agree. If volume dictates price, then knowing the origin of the volume and direction could be benefitial.

[#5]

Just add a simple MA to vol ... the example you gave using the 3 day SMA on EXX had ave of .... first day 489,705 .... 2de day 719,845 .... 3de day 1675,000

[#6]

Where did you get ave trade size?

[GarethJ]

I'd say these are two separate measures (Ave Trade Size vs. Ave Trade Direction?). Using EXX as an example: Today the final Average Trade size was 1527 (there was an individual trade at 16h48 of 281,173 - R30million!!) - yesterday was 751. There is no doubt that some serious institutional activity happened today (unless Eddie was busy again.)!! The price went up by 3% today, so we can assume that this is buying - the big increase in OBV confirms this. Yesterday was down 4% on relatively smaller trades - more panic selling amongst "retail" traders? Measuring dynamic intraday price direction (bid vs ask) appears to be much more difficult to do, but could be useful too i.e. is an individual trade initiated by buyers or sellers? To measure this you would need to measure the direction of every single trade in real time. There are a few research papers and articles discussing this on the internet, but this is impossible on OST currently.

[GarethJ]

Calculated manually - Volume/Deals. Makes it difficult to use to filter shares etc.

[SimonPB]

every trade is a buyer AND a seller .. point is the buyer is putting out risk by entering, the seller not (shorting aside) .. that's why volume is imnportant .. forget whether the trade is a buy or a sell ..

[StanleyB]

It will always be difficult to get a good handle on the trade sizes that is moving the price. One large buy order matched against 20 small sell orders is reported as 20. The auctions show as 1 trade & may also report as one.

[SimonPB]

further to my comment .. is it sellers or buyers ?? well if down then sellers and if up then buyers ..

[john_1]

agreed simon.. but what is sugested here is interesting.. trade size.. it does provide clue to the mind of the market.. I really like the idea

[SimonPB]

ya, trade size I like .. also # of trades referenced to average ..

[john_1]

yes.. over time to develop an indicator it would have to be a trade size expressed as a % of price.. because its natural to assume that pos size will decrease as price increases.. as you can afford to by less.. but the idea of who is buying as expressed by possition size is a little bit of clever!

[RichardS]

Hi Gareth, I agree with the approach. I'm not sure whether there is a scientific justification for it but volume trades (buying or selling) normally triggers a review of my strategy on a share. I sit up an take notice/ look out for a trade size 25,000 - 50, 000 on my watchlist - mainly ALSI Top 40. The trade size would normally indicate institutional activity. My thinking is that, in addition, to my own analysis, instituions would have access to a great deal more resources and insider information than I would. So, I ignore institutional trading activity at my peril. But, I don't automatically default to the trend.

[#16]

See told you someone (Simon) would come up withthe buyer AND seller comment ;-)....but the volume vs trade size could indicate a variety of options. Are the institutionals selling into strength and a tail could form in a day or teo or are institutions buying into weakness as they have that edge over retails buyers? So yes all is valid and volume is important but knowing how many bids are matched as pposed to offers could also help to refine an indicator. Also their must be a way for the JSE to track who bought/sold 500,000 shares and so on. That way we can dump the analysis of trying to match the 500,000 order covered over 100 trades. Ultimately a good idea to keep in the analysis box.

[SimonPB]

well only because it is true ..

[GarethJ]

While thinking about this, I have been troubled by the fact that I first stumbled upon "abnormal" trading activity on Brait (BAT) on 27 May last year. However, this wasn't a few big trades, but rather a non-stop barrage of small trades of exactly the same odd size (1113 or something) occurring throughout the day - certainly not "retail trading". I bought and the price shot up the next day. So while abnormal individual trade sizes may be an indicator sometimes, we also need to consider that institutions also appear to often break up large trades or use automated engines to trade. I will certainly explore this further, but for those with an appetite for 26 pages of complex maths and statistics, I have just found this. Interesting reading (well the conclusion anyway.): http://sbs-xnet.sbs.ox.ac.uk/tarun_ramadorai/TarunPapers/caughtontape.pdf

[GarethJ]

On this topic, does no-one know of a book written by an ex institutional trader that would divulge all the typical strategies they use? May help us learn to spot such activity.

[quovadis]

Price Streamer market depth (when its working) may well help in highlighting institutional activity. Look at the number of relatively small orders which are not for round numbers of shares and where the bid or ask prices change (commonly by 1c) at such a rate that one would be forgiven for thinking that the trades are computer generated.