Wrapping up Week 2
The conclusion at the end of Week 2 indicated that there appeared to be an excess of 1.2M more Buys than Sells over the week. Friday the 1st Sept showed significantly increased volume and an excess just on that day of +730K Buys over Sells – it was predicted that there would be Sells reported late for Week 2.
This predicted was borne out when a 1M trade for Friday the 1st showed up after hours on the 5th of September – it is shown marked No. 1 below. The trade appears repeated due to an error with the LSE trade log reporting – the last trade of each day appears repeated the next morning. If you look back at the trade analysis for Friday the 1st you will see that a price of 10.0625p indeed turns out to be a Sell – this 1 million Sell then brings the nett trade excess down to +200K more Buys than Sells – much closer to the balance that you would expect for Market Makers who are trying to stay neutral.
Monday 4th Sept 2017
Trading appeared to start rather slowly – with only some sells in a tight band around 10.1p showing. These were offset by a 216K buy at 8:15am that was reported exactly 2 hours later. The prices paid appeared to drop as low as 10.0p at 9:24am – but this and the subsequent 200K trade were later confirmed by a LTH as a Bed & ISA combined trade. A similar Bed & ISA trade appeared at 11:09am – also at deceptively low prices.
At 11:09am it appeared that there had been only about 100K more Sells than Buys – trading was well balanced – yet the Ask suddenly dropped. 6 minutes later at 11:15 and 11:16 there appeared sells of 100K which increased the excess of sells to approximately 300K – strange how the behaviour of the Ask appeared to presage what was to happen in the trading. With buyers able to pick up shares at only 10p until about noon Buys soon exceeded Sells – reaching an excess of 150K just before 1 ‘clock with a (delayed by 3 hours this time) Buy of 250K at 10.19p.
Things appeared to quieten down and not much happened until a quarter past 3 in the afternoon – when there is an interesting sequence of large trades, all of course reported late. At 3:15 there was a sell of 250K, followed 5 minutes later by another sell of 50K. These were rapidly absorbed by two 245K Buys at 3:22pm – and trade balance was restored 20 minutes later with another sell of 200K. So 500K sells and 490K buys all behind the scenes in less than half an hour- seems like a Big buyer may be getting some help from the Market Makers – evidently they are happy to quickly step in and take whatever the market throws at it – with no impact whatsoever on the Bid or Ask. Earlier in the day the Ask was dropped purely on an excess of 100K sells – now nothing happens when there is an excess of 300K sells – evidently the Market Makers know who to go to when they need someone to Buy shares.
At the end of the day the SP ended at 10.0p down 2.4% on the day, on a volume of 4.4M (2.3M Buys, 2.1M sells).
At this point, apart from the somewhat unusual behaviour with the Ask at 11am everything seemed to make sense, until …
Trades Reported Late
Reported at approx 2:30pm on 5th Sept – 11:11am – 519K trade at 10p (marked 3 in the log above)– this would appear to be a significant Buy. AND then another 519K trade, reported a couple of hours later, but corresponding to a trade some 21 seconds earlier than the first 519K trade (marked 2 in the log above). These must be separate trades and must surely both be Buys.
What does this do to the nett balance of trades?
Well naturally it sends the Nett trade position skywards. The only thing that could balance this, and explain the sudden drop in Ask at 11:09am is a Sell of 1M just before the two 519K Buys. Further delayed trades expected.
Tuesday 5th Sept 2017
Trading appeared to open with a Sell of 100K – this was rapidly taken up with a Buy of the same amount only three minutes later – both of these trades were reported after an hour delay – again the MMs do not wish to reveal the presence of large trades in real time. After a short period when the Ask was increased to 10.5p trading continue primarily with Buys (albeit with a Bed&ISA pair at 9:40am) until 9:53am – when there were two consecutive Buys of 100K in the space of 4 minutes. Again these trades were reported 1 hour late -what happened in real time was that by 9:59am the Bid had dropped to 9.75p – how come ?
As can be seen from the plot above Nett trades are clearly strongly favouring Buys over Sells over the whole day – even once the 400K sell at 4:29pm (reported a day late) is factored in. We may hypothesize that there is a Sell of 300-500K yet to be reported that coincides with the time that the Bid was dropped. It will be interesting to see whether this happened before the 2x 100K Buys at 9:50ish or as a result of them – this may indicate whether it is the Buyer or the Seller that is initiating the big trades. Either way it seems that a large Buyer and/or a large Seller is present but the MMs wish to obscure this fact by delaying the trade reporting so much.
Trading continued relatively slowly for the rest of the day – with Buys dominating Sells until the very end of the day when a 400K Sell was dropped in 5 seconds before trading closed.
The share price ended at 10.0p, flat on the day with a volume of 2.0M (1.1M Buys, 0.86M Sells). Further delayed trades expected.
Wednesday 6th Sept 2017
Trading really kicked off, with a couple of closely spaced 100K Sells at 8:26am and 8:27am – these were rapidly reversed within 8 minutes by big buys of 350K and 250K. A subsequent sell of 221K at 9:20am brought the nett trades figure back into the +200K/-200K band where it stayed for most of the day. As has been quite often the case in the last few weeks – there was a slight drop in the average trading price between 1 and 3pm – reflected in the drop in the Ask – however this was reversed by a 100K Buy at 3:01pm. Unlike previous days there were no Bed & ISA combined transactions identified during the day.
At 3:40pm there was another 100K Sell – and again this was rapidly taken up by a similarly sized buy within 2 minutes. It would appear that the 100k sellers are leading the action but these are being rapidly reacted to by a similarly capable Buyer or Buyers. All of this happens without any change in the Bid or the Ask – so how does the Buyer(s) know when to act ? Perhaps they are known to the market makers and are tipped off when the MMs have just got shares to sell. Or a Buyer with deep pockets is watching the level 2 screens like a hawk…
Nett trade neutrality appeared to be restored within minutes of the end of trading (4:27pm)– certainly that was how it looked on the day. However after hours trades, reported the following day, again rapidly absorbed the last minute Sell. This time it cannot be a Buyer who was sitting waiting to make a trade – instead the Market Makers appear to know of a Buyer who can respond rapidly to pick up the shares that the Seller or Sellers deposited into the market only a few minutes earlier – and they give them the opportunity to do this after normal trading has finished.
The SP finished flat at 10.0p, as it did for the previous two days, on a volume of 2.7M trades, down slightly (1.5M Buys, 1.22M Sells).
Thursday 7th Sept 2017
Trading started relatively slowly – with only small (<50K) and medium (50-100K) trades until about 11:30am. These were predominantly buys and the nett trades increased to greater than 250K Buys more than Sells. A large sell of roughly 125K at 11:28am was taken up by a PI’s Buy of 100K some 12 minutes later. Two further Buys at 1:30ish, one 250K in size, pushed the nett trades over 500K…
And then the Ask dropped – why ? With an excess of Buys of more than 500K the MMs should be increasing the prices they pay for shares, in order to get more of them, not implicitly reducing the SP by lowering the Ask. The only explanation is that there must be a significant sell yet to be reported in between 1:34pm and 2:17pm. The size ? – well it should be greater than 500K to dominate the nett trade characteristics as it would appear to have done to the Ask.
After 2:17pm there is further semi-regular buying – two Buys of 100K by a PI and numerous small Buys (the Buys of 2x 13081 shares, at 3:41pm and 3:42pm are presumably equal stakes for siblings in JISA accounts). A Sell of 100K a few minutes before the close of trading, plus another after hours Sell of 50K reduced the nett trades position to 643K more Buys than Sells at the end of the day.
The SP finished at 9.875p, down 1.2% on the day, on a trade volume of 1.63M (1.1M Buys, 494K Sells). Further delayed trade reports expected.
Friday 8th Sept 2017
Trading started relatively peacefully – Buys building up a nett surplus of a few hundred thousand shares by 9am, as has been usual in the last couple of weeks. A Sell of 160K at 9:21am knocked the nett trade balance back to nearer to neutrality. Despite no change in Bid or Ask there was then another significant period of pure buying, building the nett trade surplus to almost 600K by half past ten. Another large Buy of 150K at 10:54am and then…
The Ask is dropped, first to 9.5p and then further to 9.25p. There are, of course, no trades yet reported that would explain this. This is very similar to the situation that was experienced around 2pm on Thursday, which also does not have any Sell trades yet reported. In order to have such a large impact on the Ask, and for this to happen in two stages we can hypothesize that a similar flash short to that experienced on the 18th Aug may have been attempted in the 7 minutes after 10:55am. The total size of these Sells would have to be in the 500K+ range.
The first drop in the Ask also corresponds to a drop in the Bid which sparked off significant buying after 11am – another 250K shares were bought between 11am and 11:30am. A large Sell of 200K at 11:43 briefly knocked the nett trades back down to +750K excess Buys – but again this was cancelled out within half an hour – with another 250K excess of Buys appearing. At this point the Ask rose back to 10p and the brief period of cut-price Buying concluded. Trading was quiet for the rest of the day.
As on the 18th August it would seem that an attempt to manipulate the share price down has been quickly and comprehensively thwarted by the rapid buying of LTHs.
The share price finished at 9.625p, down 2.5% on the day – though this is primarily caused by the extreme, and asymmetric, spread currently in evidence, a result of whatever happened at 11am. The true mid-price is closer to 9.75p and the effective drop on the day only half of that reported. This was on total (currently reported) volume of 2.4M or which 1.8M are estimated as Buys and only 623K as Buys. Further delayed trade reports expected.
Summary for Week 3
|SP Change||TBP estimated Buys – Sells||Potential Shorting Activity||Comments|
|Monday 4-9-17||– 2.4%||+ 1,239,768||1M Sell||Predicted, yet to be reported|
|Tuesday 5-9-17||Flat||+ 252,878||300-500K Sell||Predicted, yet to be reported|
|Wednesday 6-9-17||Flat||+ 281,137||–||–|
|Thursday 7-9-17||– 1.2%||+ 643,175||>500K Sell||Predicted, yet to be reported|
|Friday 8-9-17||– 2.5%||+ 1,194,862||>500K Sell||Predicted, yet to be reported|
|Total||-6.1%||+ 3.6M||Approx 2.5M Sells|
Conclusion for Week 3
The trading logs for the 3rd week after the 18th Aug Flash short currently show an excess of approximately 3.6M more Buys than Sells. Detailed analysis of the trade reports indicate that we can expect to see delayed trades reported for 4 out of 5 of the days during the week. The prediction made at the end of last week, of a significant Sell for Friday the 1st was borne out by trades reported this week, so we may have some faith in the interpretative approach taken here.
The question of who may be making these Sells is harder to ascertain, as they are currently hidden from view – however the behaviour on Friday is clearly consistent with the behaviour observed on the 18th August – yet another friday where 1m shares were sold in an aggressive fashion – so we might conclude that it originates from the same shorting activity.
To total estimated short position at the end of week 2 was approximately 3M – this did not include the 1M sell on last Friday that was only reported this week. This brings the estimated short position to 4M. From the trading activity seen this week we can infer significant Sells of perhaps greater than 2.5M to be reported in the next few days. If we can ascribe these to the same shorting activity then this brings the total estimated size of the short position to 6.5M.