Hi Steffen Thanks for the write-up. Virtualeconomics isn't written by Tim Harford though, splendid fellow though he is - it's by me. Best Seamus

Congratulations James - "hyperlocalismisation" was a one-word Googlewhack until you posted it here! Also, good example about Moat; and yes, verification is going to be the big issue I think, and the solution probably sits in the reputation management space rather than with journalism.

Alec - while I am of course grateful to David for the courtesy and promptness with which he responded to my post, it doesn't alter the fact that his claims "you will not find these figures on the Oxford or Cambridge websites" and "the information that the Guardian uses is not available on the Oxford or Cambridge websites" is materially misleading. Much of that information is available on those websites; some of it is not. The former claim in particular appears in his own article immediately after four paragraphs in which he cites data which are freely available on the respective websites plus two figures, concerning admissions to individual colleges, which were indeed sourced through FOI. I concede there could be a legitimate debate as to whether he purposely intended to mislead the reader in this way and whether, therefore, it was technically a "lie" but in the absence of a clear statement from David himself that he intended no such thing my point stands. The clear impression he conveys, and presumably therefore intended to convey, is that all of the data in question were hard won when in reality only some of the break-outs by college and LEA appear to be publicly unavailable. Since those break-out data in particular have only been used so far to produce misleading assertions I think the greatest courtesy done here was by the university itself when it tried to prevent him having them.

Might be a stretch to get a start-up out of a couple of Twitter feeds but now hypothetically possible to make one into a business since you could insert the ads in there too and have the whole end to end content plus commercials experience happen on Twitter. I imagine some dumb money would do it too, just to say they were first to invest in a Twitter-only business...

Tom - not knocking what you've done with the available data, but I still contend that the available data isn't sufficient to support a conclusion either way. Re your suggestion, I have asked the university for the data I would need to do a proper analysis myself and they are looking into it, but to be fair to them it'd be a lot of work just to pull together what's needed. I have posted a follow-up detailing this and will of course post results of any further analysis I'm able to carry out as I have it.

That sounds much simpler, good idea.

Thanks for all your comments guys. Apologies for not replying sooner: I've been playing with my kids, and will be back with a detailed response when they're asleep. Briefly, thanks to Tim Harford, Julia Paolitto, Mark Hornsey and others for the additional research they've either pointed to or sent me. I'll write a follow-up later when I've had chance to read it all. Thanks also to everyone who has tweeted or commented pointing out holes in my analysis. I will come back to these in detail also but concede that my critique is not without its own limitations. Thanks also to David Lammy who has emailed me a response, which I'll post up here with my own comments shortly. Finally, lots of people found my comparison of black admissions and Portuguese admissions strange. The comparison was made solely to made a statistical point about data mining and the laws of small numbers, since Portuguese applicants seem to be the European group which in 2009 saw the lowest acceptance rate. I do not mean to imply that British educators have the same duty to Portuguese students as they do to British ones, although - without wanting to stir up a completely tangential debate - I consider that point legitimately arguable.

James- indeed. I would have liked to see the Guardian dig into the substance of these claims and see whether they stood up to a proper kicking of the tyres, not just repeated David Lammy's one-sided agenda.

I stand corrected - updated, with my thanks.

James M - very happy to do requests for my best commenter! Short answer - I think that's pretty much what newspapers were trying to do offline in the first place. (See eg here for my take, which is really Vin Crosbie's take, on where the value of newspapers really sat - it wasn't in the production but in the selection.) There's now millions of sites that try to aggregate all the best or most relevant content, either by collective editing (Digg, Reddit, Fark, 4Chan), by individual editors (BoingBoing et al) or by algorithm (Google News, Last.fm, TIOTI etc). They're all free. I don't believe there's a gap in that market for one that're more expensive than free. Marc - that's my understanding too, that like WSJ and the FT (and apparently the Australian) they;'re going hybrid. Best of both worlds approach - get the links, get the money too. Might work if they don't pretend "free" is the same as "bloody annoying reg wall" (I pretty much ignore the FT now because of that error). Also, as my friend Ilana points out, if they get a lot better at previewing the thing they're selling - retail sites are miles ahead of content in that respect. Coops - quite right, that is another challenge for ad-funded publishers, but I'm pushed to see that as a recent development. Publishers were being squeezed by networks five years ago, maybe ten.

Jonathon - yup, saw that this morning. Amazing how people seem unaware of recent news, apparently even when it has been reported at some length in the very section of the paper they edit!

You do get the feeling from CoD that we're in training for *something*, but I can't believe it's a horde of easily-shot Nazi zombies, cool though that would be.

Ah, but the Guardian is a special case - their "Corrections and Clarifications" column is my favourite bit, and if they stopped making errors they'd have to scrap it. I particularly enjoy it when they begin a correction "in this column yesterday..." because they made an erroneous correction, which happens entertainingly frequently.

Michael - interesting point, but we are talking about disrupting Opentable which caters to 14k restaurants. There are - I see from a glance at the first result on Google - about 800k restaurants in the US alone. So Opentable doesn't even handle bookings for 5%. Admittedly we don't know the crossover though. Sringaneshr - collective bargaining could work, of course, which seems to be what the original article I linked to was proposing. Rich - thanks for the link. And welcome to Twitter, I think - I seem to have been your third tweet!

Coruskate - I think our loyal reader Gareth has the solution to this one above. He suggests that the appropriate filter would be the band's fan club. I think that ticks all the boxes. The band makes more money, even if joining is a negligible fee; it can be used as the filter for tickets (maybe just for the presale if the band doesn't expect to sell out to just fan club members); and it solves the ID/reselling problem since joining handing over photos and details is less intrusive for someone joining a fanclub than someone merely buying a ticket.

James - if you don't know what I'm on about I've obviously not said it right. I'll have another go tomorrow.

Good point Anna - FB can do even more damage by targeting the senior guys.

Lloyd - I was thinking about that myself, and to my shame when I realised that lots of people would end up using the "employee, long, options" tag (£D:ELO) I started humming "Mr Blue Sky" and now can't stop... "Real" journalism is (often hilariously or illegally) bad at this, although financial journalists are usually pretty scrupulous, lots of bloggers are better and SeekingAlpha is exemplary (requires a disclosure before you can post articles). And of course the more people do it, the more we'll know not to trust those who don't. A portable disclosure tag is an even better idea though. If only I had the first idea how to code such a thing!

Rightmove is an interesting example, since it enjoys a market share that would normally be called a monopoly but that doesn't allow it anything like enough market power to effect monopoly pricing. For one thing there are numerous aggregators (Globrix, Nestoria, Zoopla, perhaps soon Google) that make that 65% number less realistic - RM gets 65% of revenues but from a user pov 100% (or thereabouts) of inventory is on more than half a dozen other sites. For another it's only really a player in sales - rentals is increasingly handled on the free sites (Craig, Gumtree). Most of all, it can't really risk raising prices to the point where advertisers defect even fairly briefly to eg Findapropery or Primelocation, because then the user experience collapses - credibility depends on carrying (or being seen to carry) all of the properties.

James, thanks for the tip about Abe, I'd missed that. Google is the best case for a monopoly here, I agree - it is the first page of the Internet for pretty much everyone. And it certainly owns more than 40% of that market!

Juliette - "commuter's paradox" is interesting! (Though is it really a paradox? Just looks like an error to me.) Dan - it's more an observation than an assumption. I've managed more people who could have worked from anywhere at all than people who needed to be in specific locations most of the time. Few of my own jobs have been location-dependent. If there is an assumption, it's one of mediocrity. Re companies paying for commuting-only - sure, there could be a special sort of Oyster ticket that only works up to 9am (say). Actually, that might prove alarmingly popular with employers as a perk - "we'll pay for your travel provided you're on time for work".

Ah, but jelly belly are the best thing ever so deserve a special exemption!

Greg - back to the good old pre-mobile Rabbit phones of the early 90s! Another idea that was arguably ahead of its time, not just rubbish.

Adam I know, a lot of people won't be moved by the offer of a free Times online sub. But consider what you have to believe for this to work. Sky loses 1m customers a year. It has to retain 30k of those for this initiative to make commercial sense. So you have to believe that one in about 30 of the conversations the retention team have with exiting customers will go a different way because there's an offer of Times online in the bundle. Alternatively, you could believe that one in about 150 of the 5m people who bounced off the new paywall will move from an existing bb provider to Sky to get access back. Alternatively you could believe that one in about 600 of the 18m UK bb customers will move tp Sky for the same reason. Now, there's a lot of potential duplication and crossover there - some people who bounced off the Times paywall will already by Sky bb customers etc. But not that many - Sky only has 14% of that market. So even if most (all) or the people you've spoken to say that they wouldn't be swayed, perhaps some other people - especially when they're already in the confusing process of changing provider or in the middle of a sales pitch - will. And it really doesn't have to be many of them for it to work.