Interdum stultus opportuna loquitur...

Friday, July 25, 2008

OzRant: Back to 200 Pickens...

Note - from June 24th 2009, this blog has migrated from Blogger to a self-hosted version. Click here to go straight there.

Every now and then, facts get in the way of stuff we think we kow. A lot of the time, those facts embarrass or undermine things that we are almost force-fed as Gospel Truth.

Think back to when you were growing up - fingerprint evidence was considered almost like Holy Writ. If a crime scene had the 'dabs' of the likely perp, he was done for.

Later, we find that as many as one in five supposed '8-point matches' are either badly performed (including instances where the matches were found using an upside down version of the accused's prints), and that far from being unique at a 13-point match level, there are sufficient duplicate 13-point matches within each database to generate false positives. That is to say, the New South Wales database might contain a 13-point match for your prints, as might the Victorian, as might the Uzbekistani or Spanish databases; this could happen even if you have never been fingerprinted.

Since the 1980s we have bee nslowly weaned off the fingerprint idea; now it's DNA (moreaccurately DNA profiles - the test is done on specific chunks of DNA). Prosecutors have a ball saying "The odds against this DNA sample incorrectly identifying the defendant are 113 billion to 1"; thus when someone is convicted based on DNA evidence, it's routinely assumed that They Dun It.

That is now pretty much out the window: while the entire genome for each of us is (most likely) unique, court DNA evidence relies on a point-match method similar to fingerprints (for DNA it's referred to as 'locus matching'). And recent findings have shown that there are FAR more matches than you would consider acceptable if you were a defendant.

For example, in Illinois, a court-mandated search of the state DNA database found over 900 pairwise 9-loci matches at the state's database of 220,000 DNA samples (excluding related parties). That is not saying that 900 DNA profiles matched a single specimen, but that there were 1800 people who were already in the system whose DNA matched at 9 points or more. That is, there were 900 Person Ai's who could have taken the rap for completely-unrelated Person Bi (where Ai and Bi were both already on record in the DNA database). There is little or no method for determining accurately how many people who are not currently on the database will be matched to someone who is. And if you can match to an unrelated party, you can be matched to crime scene evidence when you can truthfully say I Wuz Never There, Guvna.

What was the government response to this? They attempted to shut down all further research into the matter and sought to suppress the findings.

And what is more - it turns out that the whole "1 in 114 billion" number system comes, at the end of the day, from the FBI's internal estimates, which they refuse to make public.

It also turns out - if you understand conditional and unconditional probability - that having slightly less than 1 in 200 (903/220000) 9-point locus matches is perfectly consistent with a claim that there is 99.999999999% chance that a specific 13 point match can only come from a single individual. So for the thousand people who could in principle be incorrectly identified in the state of Illinois, it would be cold comfort to know that the maths was right.

And of course not everybody in the state of Illinois is on the database; there is almost a 1 in 200 chance that they will share a 9-locus match with somebody else to whom they are not related. Say, He What Actually Dun The Deed. What you have to hope for (the thing that helps make the conditional number - 1 in 114 billion or so - valid) is the small proportion of the population which is actually in the database... this reduces the probability that your DNA matches someone with a criminal record. Start growing the database and  the odds of misidentification increase.

But that's not why we're here - I just thought it was highly interesting on several levels:

  • that by-chance DNA profile matches are far more common than one would think;
  • that neither prosecutors nor juries (nor defense lawyers, for the most part) understand the difference between conditional and unconditional probability; and 
  • that the response of the State to a new hole in its armory of 'irrefutable, incontrovertible' evidence was to try and suppress it

Bear in mind, this is before any thought of prosecutorial misconduct, evidence tampering or any of the myriad misbehaviours to which law enforcement (the State's Armed Goons) are accused of being prone.

Personally I think that most of the State's Armed Goons and their adjuncts in the legal fraternity do their jobs in good faith. They are just ignorant of what they are doing - thinking that being an instrument of the State's threat of violence is not a Bad Thing. They wrongly think of themselves as being morally superior to Mafia or street-gang 'muscle', when 99.999% of the laws they enforce have no role except to force compliance on the populace.

Major Market Indices

Well, the multi-session timeframe that I had postulated yesterday (for the wearing off of the recent bounce), proved to be wildly optimistic. Not to worry...

The broad market - the All Ordinaries (XAO) - slid pretty savagely, posting a loss of 157.4 points (3.03%), finishing at 5031 points. The index hit an intraday high of 5190.5 at 10:00 am (yesterday's close - the traded high was in the 5080s at about 10:40 a.m.), while the low for the day was 5003.9 - set at 3:08 pm Sydney time. (Note to self - there might be intra-minute lows lower than those captured in the one-minute data - check on this).

Total volume traded on the ASX was 1.31 billion units, 4.4% below its 10-day average. The ASX's daily listing of all stocks included 1330 different 3-letter FPO's which traded (i.e., had non-zero trade volume). Of these, 331 issues rose, with volume in rising issues totalling 310 million units. Conversely (or even contrariwise), 714 stocks were dragged to a loss for the session, with aggregate volume traded of 889.6 million shares. [Second note to self - start including New Highs and new Lows, buffoon... why CalcStore them if they're not used? Dumbass]

Of the 494 All Ordinaries components, 120 rose while 301 fell. Volume was tilted in favour of the losers by a margin of 3.7:1, with 189.93 million shares traded in gainers while 705.31 million shares traded in the day's losers.

The Index that forms the cash basis for the SFE's Share Price Index Futures - the S&P/ASX 200 (XJO) - dipped pretty savagely,rather savagely, losing 173.6 points (3.37%), closing out the session at 4970.5 points.

Name Close +/-(%)
All Ordinaries 5031.00 -157.40 (3.0%)
ASX 50 4879.20 -187.60 (3.7%)
ASX 200 4970.50 -173.60 (3.4%)
ASX 300 4966.30 -171.00 (3.3%)
ASX Mid-Cap 50 4852.70 -97.30 (2.0%)
ASX Small Ordinaries 2846.20 -59.90 (2.1%)
ASX 20 2802.80 -120.80 (4.1%)
ASX 100 4028.40 -144.70 (3.5%)

The "heavy hitters" of the Australian market - the ASX 20 Leaders (XTL) - was taken to the woodshed somewhat, sliding 120.8 points (4.13%), closing out the session at 2802.8 points.

Among the 20 big guns, the news was pretty dismal: not a single stock registered a gain for the session. Total volume was 218.6 million units.

And what do we see - the bounce in Financials was pretty much unwound...

  • National Australia Bank Limited (NAB), -$4.21 (13.71%) to $26.49 on volume of 37.1 million shares;
  • Stockland (SGP), -$0.57 (10.12%) to $5.06 on volume of 8.6 million shares;
  • Australia And New Zealand Banking Group Limited (ANZ), -$1.6 (8.23%) to $17.85 on volume of 19.7 million shares;
  • Commonwealth Bank Of Australia (CBA), -$2.88 (6.21%) to $43.51 on volume of 14.5 million shares; and
  • Westfield Group (WDC), -$1 (5.72%) to $16.48 on volume of 12.2 million shares.

At the other end of the market-cap spectrum lie the denizens of the ASX Small Ordinaries (XSO). The small end of the market , while still dropping overall, did significantly better than its large-cap counterpart. The Small Ords slid 59.9 points (2.06%), closing out the session at 2846.2 points.

Among the stocks that make up the Small Caps index, 51 index components finished to the upside, and of the rest, 128 closed lower for the session.

The 195 stocks which make up the index traded a total of 241.83 million units: volume in the 51 gainers totalling 81.81 million shares, with trade totalling 150.91 million units in the index's 128 declining components. The major percentage gainers within the index were
  • Rubicon Japan Trust (RJT), +$0.03 (41.98%) to $0.12 on volume of 12.1 million shares;
  • Centro Properties Group (CNP), +$0.02 (7.84%) to $0.28 on volume of 7.8 million shares;
  • AED Oil Limited (AED), +$0.13 (6.13%) to $2.25 on volume of 224.6 thousand shares;
  • Perilya Limited (PEM), +$0.03 (4.63%) to $0.56 on volume of 2.7 million shares; and
  • Tower Limited (TWR), +$0.08 (4.62%) to $1.70 on volume of 51.4 thousand shares.

In the red-zone of the little-stock index, the following list represents the biggest downers (in terms of percentage decline):

  • Sylvania Resources Limited (SLV), -$0.27 (16.41%) to $1.35 on volume of 35.4 thousand shares;
  • Mincor Resources NL (MCR), -$0.19 (9.13%) to $1.89 on volume of 2.5 million shares;
  • Aquarius Platinum Limited (AQP), -$0.88 (7.57%) to $10.75 on volume of 998.3 thousand shares;
  • Lynas Corporation Limited (LYC), -$0.1 (7.39%) to $1.19 on volume of 2.5 million shares; and
  • Compass Resources NL (CMR), -$0.13 (7.1%) to $1.64 on volume of 254.7 thousand shares.

Index Changes
Code Name Close +/- % Volume
XAO All Ordinaries 5031 -157.4 -3.03 911.6m
XFL ASX 50 4879.2 -187.6 -3.7 429.4m
XJO ASX 200 4970.5 -173.6 -3.37 804.7m
XKO ASX 300 4966.3 -171 -3.33 861.6m
XMD ASX Mid-Cap 50 4852.7 -97.3 -1.97 190.4m
XSO ASX Small Ordinaries 2846.2 -59.9 -2.06 241.8m
XTL ASX 20 2802.8 -120.8 -4.13 218.6m
XTO ASX 100 4028.4 -144.7 -3.47 619.8m
Market Breadth
Advances 0 14 35 120 51 331
Declines 20 85 157 301 128 714
Advancing Volume 0m 91.6m 142.6m 189.9m 81.8m 310 million
Declining Volume 218.6m 525.9m 653.1m 705.3m 150.9m 889.6 million
GICS Industry Indices

Among the 11 industry indices, the news was universally negative: not a single sector managed to break into the "Win" column.

Since none of the industry sectors registered a gain for the session, there is no point in burdening ourselves with the internal behaviour of advancing sectors... on to the losers.

The worst-performed index for the session was Property Trusts (XPJ), which dipped 99.3 points (6.82%) to 1357.1 points. The 21 stocks which make up the index traded a total of 121.17 million units; The 17 decliners had volume traded totalling 102.53 million units, and 3 index components rose, with rising volume amounting to 17.55 million shares, The major percentage decliners within the index were

  • Stockland (SGP), -$0.57 (10.12%) to $5.06 on volume of 8.6 million shares;
  • Macquarie Countrywide Trust (MCW), -$0.08 (8.25%) to $0.89 on volume of 5.4 million shares;
  • Centro Retail (CER), -$0.02 (6.56%) to $0.28 on volume of 7.9 million shares;
  • GPT Group (GPT), -$0.12 (6.42%) to $1.75 on volume of 10.5 million shares; and
  • Macquarie Office Trust (MOF), -$0.06 (6.38%) to $0.88 on volume of 5.7 million shares.

Just missing out on the wooden spoon was Financials ex Property Trusts (XXJ), which slid 352 points (6.5%) to 5065.3 points. The 30 stocks which make up the index traded a total of 151.86 million units; The 26 decliners had volume traded totalling 149.56 million units, and 3 index components rose, with rising volume amounting to 2.3 million shares, The major percentage decliners within the index were

  • National Australia Bank Limited (NAB), -$4.21 (13.71%) to $26.49 on volume of 37.1 million shares;
  • ANZ Banking Group Limited (ANZ), -$1.6 (8.23%) to $17.85 on volume of 19.7 million shares;
  • AXA Asia Pacific Holdings Limited (AXA), -$0.36 (7.06%) to $4.74 on volume of 3.1 million shares;
  • Commonwealth Bank Of Australia (CBA), -$2.88 (6.21%) to $43.51 on volume of 14.5 million shares; and
  • Bank Of Queensland Limited. (BOQ), -$0.89 (5.74%) to $14.62 on volume of 319.8 thousand shares.

Third-to-last amongst the sector indices was Information Technology (XIJ), which slid 29.1 points (5.6%) to 490.2 points. The 2 stocks which make up the index traded a total of 1.66 million units; both fell - 

  • Computershare Limited (CPU), -$0.55 (6.04%) to $8.55 on volume of 1.4 million shares; and
  • Iress Market Technology Limited (IRE), -$0.20 (3.2%) to $6.05 on volume of 304.9 thousand shares.

Sector Indices
Code GICS Sector Close +/- % Volume
XEJ Energy 16714.4 -33.9 -0.2 70m
XUJ Utilities 5437.6 -13.9 -0.25 33m
XTJ Telecommunications 1509.2 -4.3 -0.28 33m
XHJ Healthcare 8339.5 -94 -1.11 7m
XNJ Industrials 4796.7 -74.7 -1.53 105m
XSJ Consumer Staples 7123.2 -130.6 -1.8 35m
XMJ Materials 13285.4 -252.5 -1.87 200m
XDJ Consumer Discretionary 1640.6 -47.3 -2.8 47m
XIJ Information Technology 490.2 -29.1 -5.6 2m
XXJ Financials ex Property Trusts 5065.3 -352 -6.5 152m
XPJ Property Trusts 1357.1 -99.3 -6.82 121m

All Ordinaries Major Movers

All Ords Volume Leaders
Code Name Close +/- % Volume
NAB National Australia Bank Limited 26.49 -4.21 -13.71 37.1m
OZL OZ Minerals Limited 1.98 -0.11 -5.26 35.7m
BHP BHP Billiton Limited 37.01 -0.54 -1.44 32.8m
TLS Telstra Corporation Limited. 4.38 -0.04 -0.9 28.6m
QAN Qantas Airways Limited 3.49 -0.09 -2.51 27m
All Ords Percentage Gainers
Code Name Close +/- % Volume
RJT Rubicon Japan Trust 0.12 0.03 41.98 12.1m
GPM GEO Property Group 0.29 0.05 18.37 439.3k
INP Innamincka Petroleum Limited 0.97 0.09 10.23 810.1k
TRY Troy Resources NL 1.95 0.15 8.33 36.6k
CNP Centro Properties Group 0.28 0.02 7.84 7.8m
All Ords Percentage Losers
Code Name Close +/- % Volume
SLV Sylvania Resources Limited 1.35 -0.27 -16.41 35.4k
SPH Sphere Investments Limited 2.05 -0.34 -14.23 18.1k
NAB National Australia Bank Limited 26.49 -4.21 -13.71 37.1m
SGP Stockland 5.06 -0.57 -10.12 8.6m
MCR Mincor Resources NL 1.89 -0.19 -9.13 2.5m