‘Considered to be a representative sample’
p2pnet news | P2P:- Yesterday, "TV-shows trumped movies as the most popular files on BitTorrent sites," p2pnet posted, going on, "Transformers was the most downloaded movie, with Heroes Number One on the TV list.
Our story was based on data from Mininova and they were, "considered to be a representative sample," said TorrentFreak.
Catflap, however, begs to differ, saying, "Although they say it’s a ‘representative sample’, it’s hardly representative or even close to accurate."
He goes on >>>
The stats from Mininova only show how many times the torrent file was downloaded from Mininova. It doesn’t show any stats for any completed downloads of the video.
Even the stats for seed and leech are not up-to-date, either because their system doesn’t update them more than once every 24 hours, or because the tracker the torrent is registered with doesn’t allow scraping and after some time is removed because their system considers it to be dead.
Either way, it’s not at all a fair sample of how many times the video file has been downloaded completely. The stats on Mininova have nothing to do with that info.
To illustrate why the Mininova ‘download stats’ cannot be used to even approach a ‘representative sample’ of how many times a file has been downloaded, I’ve had some of my own torrent files removed by Mininova mods for unfounded and erroneous complaints.
I’ve then had to wait for hours for a mod on IRC, or a day or two to get in touch with one in the forums to find out what the problem is. Then explain that my files aren’t against their rules and ask to be unbanned from the upload page so I could submit the torrent again.
I use trackers that do allow scraping by Mininova. One file I’ve had to submit more than once has actually been completed more than 21, 000 times although Mininova only shows that from their website the torrent file has been downloaded just 15,000 times – in its current third or fourth submission.
Before that it (the torrent file) had been grabbed several thousand times just from Mininova alone.
During the banned period my file is still being downloaded – but the torrent file is not being grabbed from Mininova because it’s been deleted. Plus, mine – as well as the majority of torrents – are posted on other listing sites and message boards.
This also doesn’t mean that every torrent file downloaded has actually been used to download the file. People forget; or they run out of space; or something better comes along. There are any number of reasons why a person hasn’t downloaded the file. There are also any number of reasons why a person has downloaded the file more than once.
One should also not forget that once a file’s been downloaded, it’s often re-registered with another tracker. I’ve seen my files ending up listed on several so-called ‘private trackers’, where presumably companies such as BigChampagne can’t monitor and track stats.
The only way I know how to see exactly how many times a file has been downloaded from a particular tracker is by putting the torrent into the Torrentspy application and checking the stats there. But again, if the tracker doesn’t allow scraping it might not register in Torrentspy.
Torrentspy can also be misleading in that since many people use multiple trackers in their torrent files, Torrentspy (I believe) scrapes only the first tracker in the list. And that tracker may not allow scraping – so there are no stats revealed.
To have a fair, ‘representative sample’ one needs to at least use an application that does show how many completes the torrent registered. Merely knowing how many times a torrent file is grabbed from a particular website doesn’t represent anything other than that. And in such cases where the torrent file has had to be resubmitted, new stats are generated and the old ones are gone for good.
Companies such as BigChampagne that do monitor torrent traffic have a more accurate and ‘representative sample’ of what’s happening.
Perhaps they would show that Transformers and Heroes were actually downloaded ten times more often than the stats provided by Mininova show.
catflap – p2pnet
Stay tuned.
Also see:
p2pnet – Transformers most pirated 2007 movie, January 2, 2008
TorrentFreak – Top 10 Most Pirated Movies and TV Shows of 2007, January 1, 2008
Use free p2pnet newsfeeds for your site. It’s really easy!
Subscribe to p2pnet.net | | rss feed: http://p2pnet.net/p2p.rss | | Mobile – http://p2pnet.net/index-wml.php
Net access blocked by government restrictions? Use Psiphon from the Citizen Lab at the University of Toronto. Go here for details. Download here.
January 3rd, 2008 at 11:01 am
BigChampagne sees all and knows all? I don’t believe it. It’s conceivable that they could monitor some (but not all) torrent traffic. But how in the world could they monitor traffic on distributed networks without central servers, using something like kademlia protocol? I know that they claim to be able to do this, and it is in their interest to make that claim, but it doesn’t seem possible.
Would someone more knowledgeable than I am care to comment? (Almost anyone is more knowledgeable than I am.)
January 3rd, 2008 at 11:30 am
My guess is that they use scripts to find the top files for each public tracker, then sum and filter all the numbers.