You are right. Alleles were coded in 0/2 format in the plink .raw copy I was working with, and I divided with 2 to account for the two lines of descent from the common ancestor, but not with another 2 to account for the 0/2 format. This means that the age estimates posted in the blog  ought to be halved. This would make "Root" about 80ky, and CT about 31.5ky. These appear quite different from previous estimates, including the ones I originally got 
Looking at the original experiment  where I did divide with 2 twice, it seems that there are many more variant sites per pair of chromosomes; this is why the "Root" estimates for the two experiments end up similar.
Regressing the pairwise divergences of  on , I obtain:
Number of variant sites per pair of chromosomes using  = 215+ 0.07 * Number of variant sites per pair of chromosomes using 
So, it does appear that there is an excess of variant sites on average in  compared to .
Unfortunately, there is very little meta-data on the 1000 Genomes ftp site, so I don't know what to trust. If the phase1 data are to be trusted, then age estimates will appear quite smaller than those published by Karafet and Cruciani; but, perhaps, there was a little aggressive pruning of variants between the working data and the phase1 data? I guess we'll have to wait until the 1000 Genomes Project publishes on the topic to see which is the case. Thankfully, it seems that a lot of people are working on the Y-chromosome phylogeny , so I'm sure we'll find out more soon.
 http://dienekes.blogspot.com/2012/08/dates-of-major-clades-of-y-chromosome.html, using: http://ftp.1000genomes.ebi.ac.uk/vol1/ftp/phase1/analysis_results/integrated_call_sets/  http://dienekes.blogspot.com/2012/07/estimating-age-of-y-chromosome-adam.html, using: http://ftp.1000genomes.ebi.ac.uk/vol1/ftp/technical/working/20111123_chrY_snps/  http://dienekes.blogspot.gr/2012/08/ashg-2012-title-imputation-part-ii.html http://dna2012.gerichtsmedizin.at/files/DNA_in_Forensics_2012_program.pdf
So, while these calculations are valid in terms of relative ages of haplogroups, we'll have to wait for a formal publication that explains the variant filtering process better until we can get some good absolute estimates.
This represents my final take on the dating of major clades of the Y-chromosome phylogeny using 1000 Genomes SNPs from the phase1 dataset. When new data are added to the project, more haplogroups will turn up and the estimates will probably improve. Some of these may be slightly different than the original ones I posted, since there was a slight bug in the code I was using, and not all pairs between clades were compared.
I have also added the sample standard deviations for the age estimates; these do not take into account uncertainty regarding the mutation rate. Overall, it seems to me that with more samples, deep sequencing of the Y (as opposed to the low coverage of the 1000G Project), sequencing of larger regions of the Y, as well as a better estimate of the mutation rate (which can be achieved when many father-son full genomes become available), we are going to have a pretty crisp estimate of the ages of all the branching points of the tree.
- The greatest contrast in the human species is between the BT clade which is less than half the age of the most basal clade within this dataset; there are even more basal Y-chromosome lineages in African hunter-gatherers
- Most modern humans, including most Africans, belong to the BT haplogroup (~70k years). One can probably correlate the emergence of this haplogroup with a variety of archaeological events: arid periods in the Sahara and Arabian deserts, and the Toba eruption. The geographical origin of BT is currently unknown; the recent discovery of B haplogroup chromosomes in Iran and Afghanistan ought to give us pause about automatically assuming an African origin; the same can also be applied to Y*(xBT) chromosomes that also exist outside Africa in unusual places. Such chromosomes may represent recent gene flow and/or relics of great antiquity: they should be made the object of a study that will compare them to their African relatives.
- Marginally younger is the CT haplogroup, that includes the CF (Eurasian) subclade and the DE subclade (Afrasian). All of these have an age of around 60k years. If an ecological calamity is responsible for the coalescence of human Y-chromosomes at around this time, then it is easy to imagine how subsets of modern humans from the original homeland may have acquired their own unique clades. The CF clade may correspond to people who went to the Levant or Central Asia, from which they would launch the Upper Paleolithic, while the DE clade may correspond to people who stayed in the South, heading west (E) or east (D).
- The Upper Paleolithic/Lower Stone Age transition was probably made by these people 10-20 thousand years after they split up; I would wager that they already had the necessary mental "hardware" for the task, and the stress of the ecological calamity that befell them spurred them to develop the necessary "software" as well, and so did the fact that they found themselves in unfamiliar surroundings and had to quickly adapt. Hence, the appearance of full behavioral modernity sensu Klein. The major lineages that continue to dominate most humans (IJK and E) trace their origin to this crucial transitional period (50k years).
- The West Eurasian clade IJ seems to have emerged earlier (37k years) than the East Eurasian clade NO (33k years) and the Central Asian-Siberian (?) clade P (also 33k years). This may be suggestive of the route taken by modern humans in Eurasia as they moved to the east.
- There seems to be a plethora of important lineages that coalesce around the Last Glacial Maximum (I, J, O, R1). Perhaps this too was a stressful period for humans resulting in a loss of diversity and the survival of a few lineages from the early arrivals. In the case of haplogroup I which is truly European in distribution it is tempting to imagine that it represents the emergence of the Gravettians, while its sibling J stayed in the Near East, where IJ* chromosomes have recently been found.