Hi All,
Newbie to T2W, so nice to hopefully be of service. This is a question that is often banded about, but basically the reporting of trade direction in the UK is just a series of guesses, particularly when it comes to the SETS stocks. Sites like ADVFN, III etc are just working on the 'likely' direction, but their best guesses are flawed - seriously flawed. Here is a post i've nicked from 'another place' which sums most of it up.
There are several problems with all of these discussions about why has the price gone down since there were x million buys.
The first main problem is that trades are very often mis-identified as buys or sells by ADVFN and other sites, and unfortunately it is the big trades which stand the most chance of being allocated wrongly, since they are often delayed. Clearly if the odd few million shares fall into the wrong category then already the discussion is doomed. Some sites allocate them better than others, and the Exact Futures link on the trades page is the best I've seen, but at the end of the day it is impossible for anyone to be accurate about buys and sells since as harvester says the LSE does not report the allocation nor even the spread at the time of the trade.
The issues that cause mis-allocation are (a) trades of more than 6 x NMS which are delayed by one or more hours depending on NMS, so when the trade is received it is matched with the current (wrong) spread (b) single protected transactions which could have been placed at any time during the day and it is therefore not normally possible to deduce whether they are buys or sells (c) the fact that spread changes are sent immediately from the LSE whereas MMs have a few minutes to report trades, so when trades arrive the spread may already have changed (d) the fact that "buys and sells" really only has meaning if the other side of the trade is a market maker since what you are trying to do is in effect to establish the MM's book, it is illogical to classify cross trades such as M, X or AT codes since they are not trades between an investor and a market maker. For SETS stocks two further issues come into play, (e) for some stocks the spread is very small and changes frequently so any slight delay in the reporting of trades causes confusion and (f) the fact that AT trades on SETS are not trades with a market maker, they are trades where one party buys and the other sells and none, one, or both of the parties could be market makers, so it is nonsensical to allocate them as buys and sells.
The second main problem is the assumption is that price movements are dependent on the relative proportion of buys and sells. That is not the case. We pay market makers to make a market for shares, and to do that they need to be able to set any price they fancy (subject to market conditions). If they did not have this freedom, there would be cases where they would be unable to balance their books and would end up holding large long or short positions which they were unable to close. So MMs will for instance very often move prices down if there has been a lot of buying. Why? Because they need to encourage profit taking to get some of those shares back. This of course is called "shaking the trees" or manipulation by some, but it is the only possible way that a market can work. If you think of the ideal share, which is so brilliant that everyone wants to buy it and no one wants to sell it, after a very short time MMs would have no more shares to supply to investors and trading would stop, unless of course they are allowed to encourage profit taking by lowering the price.
MMs will often mark shares up or down on announcements when there has been no trading. Why did the price drop when there were no sales? Because the MMs were anticipating sentiment changing.
The third main problem is SETS, which is a totally different way of working to market makers. There are no market makers under SETS, or at least they are not called that, since it is SETS that determines the current spread. Small private investors cannot normally deal on SETS (exception: GNI) and instead deal with "retail service providers" who have to quote a spread which is at least as good as the current SETS spread. RSP trades go through with the normal MM codes such as O. SETS trades go through as AT trades, which are both buys and sells at the same time.
The way that SETS works is that buyers and sellers have two options. They can either request to deal immediately, or wait for the desired price to come up (i.e. specify a limit). Limit trades go onto the order book on the buy or sell side as appropriate, i.e. if an investor wishes to buy his offer will go on the right hand or SELL side of the book and vice versa. Those investors who wish to deal immediately instead get the top price on the order book, so an investor wanting to buy would be dealt against the left hand or BUY side of the book. So the two sides of the book never come together since the top trades are picked off as buyers and sellers arrive wishing to deal immediately.
Anyone using SETS can place offers to buy and sell on it, and the best offer to buy and sell form the spread. Users of SETS can include private investors, institutions, brokers, and market makers, and a SETS trade can happen between any of these categories.
So with SETS the buy/sell idea is totally meaningless. There is no central group of market makers which can be said to have bought more shares then they have sold today or vice versa. Instead, the spread is dependent on people putting bids and offers up on a public order book, and if sentiment is such that in the afternoon people key lower prices in, well, you could have had 100% buys in the morning but the price will still drop.
So apologies for the long post but basically the answer to this and all similar threads is that looking at trades in this way is pretty pointless, and totally futile in the case of SETS stocks.
