Last Questions

[kirbyzzz]

Hi!
I think I've gotten the hang of using Bakabt and most of the downloading/seeding parts.
I've just got a few questions left and I'll be set for life ^^
1.) What is initial seeding? Is it better than just regular seeding? I've googled it but none of the answers satisfied me.
2.) What does it mean for a tracker to be offline (timed out) and why does it happen?
3.) Is there a limit of how many things can be seeded at one time for Bakabt or through uTorrent?
Thanks in advance!

[zherok]

1. Initial seeding is for when there's only a single seed in the swarm. If you're not the initial seeder don't bother turning it on, it won't help your seeding performance.
2. The tracker is no longer connecting you to other users and in the case of private trackers like bakabt you're no longer having your stats tracked for any peers you're connected to while the tracker remains offline. They go offline for a variety of reasons, but most frequently it goes down for maintenance at a set time and lasts about an hour.
3. Nope, though you can limit it yourself if you wanted to through your client. Still capped by your upload limit though.

[Al_Sleeper]

1. Initial seeding is for when there's only a single seed in the swarm. If you're not the initial seeder don't bother turning it on, it won't help your seeding performance.

Besides, from my experience, its net effect is close to zero on BakaBT.

[zherok]

Apparently different clients implement it differently. But are there that many torrents here that even have a single seed for all that long?

[Van.]

But are there that many torrents here that even have a single seed for all that long?

Think about the meaning of the word "Initial" 
This function is only for the initial seeding, when only one peer has all files.
I think it's necessary only for seeding of large torrents on a slow connection.
I personally never used it.

[Bozobub]

{snip}
3.) Is there a limit of how many things can be seeded at one time for Bakabt or through uTorrent?
Thanks in advance!

You can find several useful settings re: your torrent queue in uTorrent at Options>Preferences>Queuing.

"Initial seeding", BTW, is meant for when there are *very few* seeds, not necessarily only one.  It has the most effect for a single seed, but still is of some use for very small ratios of seeds:peers.  The usefulness of initial seeding, however, drops off VERY quickly as the seeds:peers ratio increases.

[Freedom Kira]

There was a guy on another forum who decided one day to test initial seeding performance on a brand new torrent on different clients. It turns out that BitTornado's Super Seeding mode is more efficient than uTorrent's Initial Seeding mode (both 1.8.5 and 3.0.x). Using BitTornado, more seeders showed up after he seeded about 1.2x the torrent size. In uTorrent, it wasn't until about 1.5-2x the torrent size. The downside to BitTornado is that it hasn't been updated for years (perhaps that's a good thing for those who absolutely hate updating software).

Just a bit of fun trivia to add to your knowledge about Super Seeding.

[zherok]

Think about the meaning of the word "Initial" 
This function is only for the initial seeding, when only one peer has all files.
I think it's necessary only for seeding of large torrents on a slow connection.
I personally never used it.

That was my point. But I'm talking about bakabt in particular: an archival tracker with an occasionally lengthy approval process and files that are frequently easily available elsewhere (especially newer anime) and members who actually compete to get offers up.

Probably more useful on those gigantic FLAC collections or something though.

The usefulness of initial seeding, however, drops off VERY quickly as the seeds:peers ratio increases.

That was my thinking. The newest anime release to go up on the tracker (added today) already has 156 seeds. The window for initial seeding is ridiculously small here.

[Bob2004]

There was a guy on another forum who decided one day to test initial seeding performance on a brand new torrent on different clients. It turns out that BitTornado's Super Seeding mode is more efficient than uTorrent's Initial Seeding mode (both 1.8.5 and 3.0.x). Using BitTornado, more seeders showed up after he seeded about 1.2x the torrent size. In uTorrent, it wasn't until about 1.5-2x the torrent size. The downside to BitTornado is that it hasn't been updated for years (perhaps that's a good thing for those who absolutely hate updating software).

Just a bit of fun trivia to add to your knowledge about Super Seeding.

I was always under the impression that, rather than trying to get more seeders (ie. people with the entire torrent) as quickly as possible, the point of initial seeding was to get different bits to as many different peers as possible. In other words, try and get all of file 1 to one person, all of file 2 to a second, file 3 to a third, etc, so that the entire torrent is available from sources other than the initial seeder as soon as possible (and so that the load from other leechers is spread out as quickly as possible, too, since they can all download from each other).

So measuring the time it takes for the first other person to get the entire file is a fairly useless way of measuring efficiency, since it's a totally irrelevant statistic - it relies on the efficiency of the first few peers who get all the different bits of the torrent as much as, if not more than, the initial seeder.

Of course, I could be wrong, but that's how I understand it to work. And it does make a lot more sense, IMO.

[Bozobub]

You are quite correct, Bob2004.

[megido-rev.M]

I was always under the impression that, rather than trying to get more seeders (ie. people with the entire torrent) as quickly as possible, the point of initial seeding was to get different bits to as many different peers as possible. In other words, try and get all of file 1 to one person, all of file 2 to a second, file 3 to a third, etc, so that the entire torrent is available from sources other than the initial seeder as soon as possible (and so that the load from other leechers is spread out as quickly as possible, too, since they can all download from each other).

This is definitely the case for utorrent.

[Freedom Kira]

That is quite correct. You are also correct that it is useless to judge based on time, which is why it is not measured based on time, but rather how much of the file the initial seeder must seed before it is fully distributed.

I can't remember if he was waiting until more seeders showed up or if he was waiting until the swarm availability reached 1.0. The latter case would probably be more meaningful.

[Bob2004]

That is quite correct. You are also correct that it is useless to judge based on time, which is why it is not measured based on time, but rather how much of the file the initial seeder must seed before it is fully distributed.

I can't remember if he was waiting until more seeders showed up or if he was waiting until the swarm availability reached 1.0. The latter case would probably be more meaningful.

Ah, OK. Your post made it sound like he was measuring how long it took until another person had downloaded the entire file (and became a seeder rather than a leecher). Which I'd argue is not a useful statistic to measure - but yeah, swarm availability is much more logical.