I've entered some 160 trackers into Settings/Transfers/Trackers/Individual Tracker Options/Add To New to hopefully increase my upload and download volumes per unit time for my every individual transfer when I'm using public trackers only.  Many of these are located across the globe, many are limited, good for some sorts of content but not others, good at some hours but not others, and so on.  Without any refining, this arrangement increases both my upload and download rates by 50-100%.  Typically 25-40 contributing trackers remain after a transfer's rate/volume stabilizes, as much as it ever does.

If I could, I would sort the list to put the most suitable trackers at the top, so that Tixati would stick with those "better" ones for the duration of the transfer and never even get to the "worse" ones toward the bottom of the list.  It's possible that's impossible, however, and is certainly impractical to do so manually.  Does Tixati allow a user to programmatically evaluate the responses returned by Tixati for each tracker it attempts to connect with?  (Is a comprehensive list of return codes available?)  If Tixati did provide for such, I could cull the list while a transfer runs, deleting trackers that fail (e.g.,"Connection Refused", "HTTP 403", "Invalid Response", "Host Not found", "Tracker failed...", "Bad SSL Data") or perform worse than I desire (e.g., "Retry in 30 minutes") as Tixati goes down the list trying them.

Maybe there's a better way to achieve my desired end.  I'm ignorant of most of the details held in the socket parameters section (Network/Connections/Advanced Socket Options).  Some of them look enticing to my naive or ignorant eyes.  Thanks for your suggestions.

160 trackers per torrent? that is insane. Adding more trackers then what a torrent comes with already is rarely useful, because every person getting that torrent would have to add them as well, and largely unneeded with DHT.

I highly doubt this will gain any traction.

I used to do this for a while, adding a mountain of custom trackers to all xfers by default. I found it actually a slower/worse pickup than pure DHT as a great many of the trackers had trolls monitoring/etc. so in short order I'm getting 200kb/sec connected to 300 fake MAFIAA peers until it slowly weeded those out for more speed (15mb/sec is closer to home). With torrent trackers+pure DHT the speed and valid peers pickup more quickly..

My hypothesis was that a torrent might be presented with fewer trackers than are actually serving the torrent, and so in adding more trackers, I'd connect to more peers.

You say this is obviated by DHT?  What is it about DHT design or operation that makes it unnecessary for a tracker to mediate between peers?  Are trackers even necessary under DHT once the transfer is running?  Does DHT maintain a pool of available peers and file segments?  (That seems unlikely since it flies against the decentralization of p2p, but, as I said, I don't know how it works.)


I didn't consider MAFIAA.  Good you mention them.  I did take about 30 of what seemed to me to be the top trackers and used just them, rather than keep the whole ~200 tracker list.

This seems to work OK, however, I admit I did no quantitative testing or comparisons.  I'd like to be able to instrument my copy of Tixati or use an instrumented copy to compile some meaningful test results.  And, still, with the ability to dynamically prune for every individual transfer the trackers that are useless, I could overcome that problem.

Until I learn how p2p and DHT work, I'm limited to speculation and reliance on developers who understand them.

Trackers are not even needed when the Torrent is first loaded. You can have torrent with no trackers at all.

First tracker do not track file segments each peer/seed has. They ONLY track peer/seed IP/Port, and amount Uploaded/Downloaded, and Torrent Client Name/Version.

DHT (Distributed Hash Table) + PEX (Peer exchange) make trackers for public torrent unnecessary. They can still help a little of course but little over what those two can do on their own.

So my idea amounts to a lot of work for little value?  OK.  What can you suggest for wringing more data from the seeds via DHT or PEX?  Thanks for your explanations and comments.