"Jumping on", overseeding and related stuff
-
How about some limitation on seeding torrents acquired elsewhere but posted here by someone else?
For example - look at: http://tracker.gaytorrent.ru/details.php?id=47070&dllist=1#seeders
At the time I viewed it 16 of the 26 seeders had 0.0 kb downloaded for that torrent, meaning they got it from somewhere else.
Look at the ratios for the people who did download it from here or are in the process of downloading. There is very little liklihood they will ever be able to get a decent ratio for that torrent.
I have no problem with the original torrent poster having 0.0 because he took the time to share it here, but the others are just jumping in to inflate their own ratios at the expense of others who might need to upload.
How about having the tracker reject all such attempts if the person's ratio is over 1.0 and he is not the poster? Give the rest of us a chance to give back what we download.
At least limit the number of such active "seeders" after there are actual completions here.
I wonder how many power users got their status this way?
TR
-
i can please you a bit, the problem you mention is related to overseeding, and is already in discussion for a longer time. however, one precision has to be said: our tracker measures only "on the fly" values, aka "per session", it means that once you restart your connection or your ISP disconnects you even for a short time, your values are reset to zeroes, therefore this is not a good way to track possible (let's call them) "external seeders". the tracker software needs to get significantly reworked to track such events, currently it's technically not possible.
-
–-
While the tracker may not show it accurately on the seeders/leechers page for a torrent, I think you will find that it does show the DL value accurately for each individual torrent on the user's own page.
TR
-
As well session based in the per torrent display. Only the total ip-/downloads are accumulated.
-
How about some limitation on seeding torrents acquired elsewhere but posted here by someone else?
For example - look at: http://tracker.gaytorrent.ru/details.php?id=47070&dllist=1#seeders
At the time I viewed it 16 of the 26 seeders had 0.0 kb downloaded for that torrent, meaning they got it from somewhere else.I get your point, in that if they DL it elsewhere, then it is counting against them on the other site and not here. I doubt this is really that big a problem, though. Just a hunch, not proven.
At the time I viewed it 16 of the 26 seeders had 0.0 kb downloaded for that torrent, meaning they got it from somewhere else.
As already mentioned, that's not a good way to tell if they got it here. More correct, but tedious, would be to compare the names of the users in the "Peers" list (you can click for full list) with the users listed in "Snatch Details" at the bottom of the torrent page. While some users have anonymous settings so not 100% accurate, it would exonerate someone who has a 0.0 DL currently (for reasons listed by mods here) but had, in fact, downloaded the file at this site.
I have no problem with the original torrent poster having 0.0 because he took the time to share it here, but the others are just jumping in to inflate their own ratios at the expense of others who might need to upload.
How about having the tracker reject all such attempts if the person's ratio is over 1.0 and he is not the poster? Give the rest of us a chance to give back what we download.
At least limit the number of such active "seeders" after there are actual completions here.
I think this kind of jump-on seeding would be hard to police. Regardless of the technical problems (which I leave to the more knowledgeable), the question of fairness isn't as cut and dry as you say. As example from my personal experience: I ripped one of my DVD's and posted it on another site. I was going to post here next, as soon as I had seeded enough on other site. When I got here, someone else who had DLed it at the other site had posted it here before me. So I added myself to the seeding pile, with a 0.0 DL of the file here. In your scenario that makes me bad, but I was the original ripper and poster. I am not complaining; I have posted files DLed from other sites occasionally as well. Just pointing out that things are rarely as simply as they can seem.
I wonder how many power users got their status this way?
Not many. It seems unfair, I know, but it just doesn't logically gain that much. If they got it elsewhere, they probably took a hit to their ratio there (if the elsewhere has ratios). So they couldn't play the transference game long. And its a lot of trouble to go to on any sort of sustained basis. My guess is most Power Users got that way through uploading new torrents and seeding fairly, and probably have an average-or-better upload speed. Getting to PU isn't that hard; the ratio is only 1.05 and some time on the site. It really doesn't take all sorts of cross-torrent machinations.
-
I think this kind of jump-on seeding would be hard to police. Regardless of the technical problems (which I leave to the more knowledgeable), the question of fairness isn't as cut and dry as you say. As example from my personal experience: I ripped one of my DVD's and posted it on another site. I was going to post here next, as soon as I had seeded enough on other site. When I got here, someone else who had DLed it at the other site had posted it here before me. So I added myself to the seeding pile, with a 0.0 DL of the file here. In your scenario that makes me bad, but I was the original ripper and poster. I am not complaining; I have posted files DLed from other sites occasionally as well. Just pointing out that things are rarely as simply as they can seem.
You do make a good point as the original ripper. But my question is why didn't you post to both at the same time? That would have avoided the problem. I have seen quite a few who do that.
… If they got it elsewhere, they probably took a hit to their ratio there (if the elsewhere has ratios). So they couldn't play the transference game long. And its a lot of trouble to go to on any sort of sustained basis.
They could have continued seeding at the other site, while also jumping on the post here.
Also, whoever posted the torrent here has their leech/seed connections number limited. These additional seeders are not so restricted. They could have any number of these going.
And I have some evidence that it is a bigger factor than you might imagine.
TR
TR
-
…
Also, whoever posted the torrent here has their leech/seed connections number limited. These additional seeders are not so restricted. They could have any number of these going.And I have some evidence that it is a bigger factor than you might imagine.
...Having the number of seed [nb]typical 1[/nb] and leech connections [nb]typical 0 to 5 with ADSL copper based internet access[/nb] limited does not limit your upload to my experience. A new upload will take all your bandwidth with a potentail to go over 1 MBytes/s, so on a ADSL service with ~128 KBytes/s it does not limit. When it is new and another seeder jumping on with higher upload will leave nevertheless enought for the original poster to run his single allowed upload full rate. There are always soooo many leecher with better download than upload. Remember, as soon as there are 3 other seeds, it's an "old" torrent and it's free to upload or leech more.
[off-topic]When people posted my original rips and posts cross site, without giving credits ("not my rip" style), I was jumping on them and pushing my full upload bandwidth into it, because it displeased me. Now I have learned to live with the fact that once you upload a torrent, you lost it's control and file shring being not fair. My last new torrent posts I run with 150 kBytes/s only [nb]instead of full rate 1 to 2 MBytes/s[/nb], still I get up easly 4x the file size before the buzz is over[/off-topic]
-
You do make a good point as the original ripper. But my question is why didn't you post to both at the same time? That would have avoided the problem. I have seen quite a few who do that.
I am more likely to do that now, but it depends on where I upload it from. I have access to a high speed line sometimes, but normal home line is slower, like most people. Seeding a file on two site simultaneously splits the speed between them, causing the whole process to take longer to get a file seeded. If it was just about me, I guess it wouldn't matter. But I try to get the file out as fast as I can, to get the torrent cloud moving.
I now feel as Uwe does. Once I put it out there, it's not "mine" anymore.
And I have some evidence that it is a bigger factor than you might imagine.
It seems there would be two issues.
- How much does this jump-seeding occur?
My inclination was not much, but if you have evidence otherwise, it would be interesting to see. - How much does it really affect the ratio of others?
My rough mental calculations would suggest that a couple of jump-seeders wouldn't be much of a factor on a torrent with lots of leechers, unless they made up a significant percentage of the cloud. All things dissipate over large numbers. A jump-seeder is only contributing one extra copy to the cloud. If they seed to a ratio of 1.0 or so, not much impact. Far more likely of a problem, causing others difficulty getting to a higher ratio, is power-seeding, regardless if they got the file here or elsewhere. Power-seeders with strong upload ability can contribute many more copies to the cloud, which is really what throws off a ratio. There is another thread elsewhere in the forum where power-seeding is discussed more in depth.
- How much does this jump-seeding occur?
-
And I have some evidence that it is a bigger factor than you might imagine.
It seems there would be two issues.
- How much does this jump-seeding occur?
My inclination was not much, but if you have evidence otherwise, it would be interesting to see. - How much does it really affect the ratio of others?
My rough mental calculations would suggest that a couple of jump-seeders wouldn't be much of a factor on a torrent with lots of leechers, unless they made up a significant percentage of the cloud. All things dissipate over large numbers. A jump-seeder is only contributing one extra copy to the cloud. If they seed to a ratio of 1.0 or so, not much impact. Far more likely of a problem, causing others difficulty getting to a higher ratio, is power-seeding, regardless if they got the file here or elsewhere. Power-seeders with strong upload ability can contribute many more copies to the cloud, which is really what throws off a ratio. There is another thread elsewhere in the forum where power-seeding is discussed more in depth.
I will use a torrent you uploaded as an example. (#47070)
Within a short time of you posting the torrent there were 11 seeders. I would guess that no more than 1 or 2 downloaded from here.
Anyone who does download from here will have to contend with the others to upload a significant ratio of what they have acquired here. It is not just your torrent, it is very widespread. And it is not just a matter of 1 or 2 jumping on.
If you want to test this, upload a torrent from another site and watch how many completions there are before you are anywhere near 1.0 on your torrent.
More complete info will be coming later (next month).
TR
- How much does this jump-seeding occur?
-
I will use a torrent you uploaded as an example. (#47070)
Within a short time of you posting the torrent there were 11 seeders. I would guess that no more than 1 or 2 downloaded from here.
Anyone who does download from here will have to contend with the others to upload a significant ratio of what they have acquired here. It is not just your torrent, it is very widespread. And it is not just a matter of 1 or 2 jumping on.
If you want to test this, upload a torrent from another site and watch how many completions there are before you are anywhere near 1.0 on your torrent.
TR
According to the more detailed info ThrobbinRod sent me in a private message, the 11 seeders corresponded to a snatched amount of 11 times, indicating no jump-seeding likely.
Just so everyone reading is clear: the number of completions is exactly that - completed downloads from the site. If someone gets a file from another site and jumps on to a torrent here, they get added to the seeder total only, not the completed total.
However, ThrobbinRod is correct in that a jump-seeder can take upload away from the initial seeder of a torrent. He and I are arguing about how much this really happens in practice, but it is certainly possible in principle.
-
Here are some stats for us, especially ThrobbinRod, to ponder over. Not sure if it supports you or not.
Last night I uploaded a torrent for a Sk8erboy film, Boycatcher. The original rip was not my own, but from another popular site. It took about 11 hours to spread. Because my upload speed was only about 60kb/sec, others in the torrent could spread faster. When I looked this morning, there were dozens of people in the 95%-99% completed range, just about to cascade over into seeder. I was the only seeder still. I was curious, so I watched and refreshed the torrent page to see it happen.
Here are the current seeder stats, taken only about 20 minutes after the cascade of seeders happened:
43 Seeder(s)
User Connectable Uploaded Upspeed Downloaded Downspeed Ratio Completed
(Anonymous) Yes 11.59 GB 125.54 KB/s 1.28 GB 7.64 KB/s 9.065 100.00%
gayron No 6.66 GB 158.23 KB/s 1.28 GB 28.21 KB/s 5.208 100.00%
xpanda Yes 6.45 GB 358.42 KB/s 1.28 GB 31.79 KB/s 5.046 100.00%
chita2 Yes 5.84 GB 162.94 KB/s 1.28 GB 32.44 KB/s 4.566 100.00%
xeseerf6699 Yes 4.76 GB 187.7 KB/s 1.28 GB 30.47 KB/s 3.721 100.00%
bruetzel123 Yes 3.19 GB 68.93 KB/s 1.28 GB 25.35 KB/s 2.497 100.00%
mifka1902 Yes 2.66 GB 599.89 KB/s 1.28 GB 37.46 KB/s 2.078 100.00%
geekguy Yes 2.12 GB 60.15 KB/s 0.00 kB 0 B/s Inf. 100.00%
adrianboy Yes 1.80 GB 43.05 KB/s 1.28 GB 31.97 KB/s 1.408 100.00%
alp No 1.69 GB 517.2 KB/s 1.28 GB 27.12 KB/s 1.323 100.00%
cockring1000 No 1.59 GB 54.77 KB/s 1.28 GB 33.24 KB/s 1.244 100.00%
whatalife Yes 1.48 GB 77.33 KB/s 872.19 MB 5.52 KB/s 1.739 100.00%
Bangboy Yes 1.34 GB 119.05 KB/s 1.25 GB 32.35 KB/s 1.073 100.00%
hucul2001 Yes 1.20 GB 44.31 KB/s 1.28 GB 33.76 KB/s 0.936 100.00%
sexydude Yes 1.10 GB 68.12 KB/s 1.28 GB 26.32 KB/s 0.859 100.00%
snooze9 No 1.05 GB 40.31 KB/s 1.28 GB 27.78 KB/s 0.821 100.00%
Zappy Yes 1.03 GB 43.7 KB/s 1.28 GB 27.78 KB/s 0.804 100.00%
edmondo No 918.47 MB 68.85 KB/s 1.28 GB 34.19 KB/s 0.701 100.00%
schlagzeug Yes 700.30 MB 36.18 KB/s 1.28 GB 30.05 KB/s 0.534 100.00%
swat0 Yes 684.83 MB 25.64 KB/s 1.28 GB 32.7 KB/s 0.523 100.00%
rapid266 No 540.39 MB 20.75 KB/s 1.28 GB 32.67 KB/s 0.412 100.00%
martin1985 Yes 536.90 MB 4.15 KB/s 1.28 GB 29.77 KB/s 0.410 100.00%
tutschi186 Yes 534.67 MB 27.5 KB/s 1.28 GB 29.5 KB/s 0.408 100.00%
SebiJan Yes 480.42 MB 24.41 KB/s 1.28 GB 20.88 KB/s 0.366 100.00%
bryan90 Yes 458.70 MB 4.7 KB/s 1.28 GB 26.58 KB/s 0.350 100.00%
hotdawg Yes 436.85 MB 37.79 KB/s 1.28 GB 30.62 KB/s 0.333 100.00%
amoazotador Yes 435.66 MB 0 B/s 580.00 MB 0 B/s 0.751 100.00%
molleke No 295.58 MB 11.66 KB/s 1.19 GB 28.93 KB/s 0.242 100.00%
iceboxcool Yes 242.69 MB 12.44 KB/s 1.28 GB 36.76 KB/s 0.185 100.00%
riverghandi Yes 232.59 MB 124.86 KB/s 780.00 MB 32 KB/s 0.298 100.00%
serfo No 226.85 MB 21.23 KB/s 1.28 GB 257.44 KB/s 0.173 100.00%
loadjunkie Yes 218.66 MB 15.85 KB/s 1.28 GB 103.32 KB/s 0.167 100.00%
gaytrekkie Yes 199.06 MB 53.73 KB/s 1.22 GB 31.11 KB/s 0.159 100.00%
hjelm42 Yes 122.56 MB 5.28 KB/s 1.28 GB 34.03 KB/s 0.093 100.00%
(Anonymous) Yes 111.96 MB 5.06 KB/s 1.26 GB 29.4 KB/s 0.087 100.00%
(Anonymous) Yes 106.30 MB 2.38 KB/s 1.28 GB 21.98 KB/s 0.081 100.00%
Andryppp No 83.48 MB 27.84 KB/s 84.19 MB 26.02 KB/s 0.991 100.00%
tigergutt Yes 56.60 MB 1.82 KB/s 1.28 GB 35.67 KB/s 0.043 100.00%
vinoario Yes 26.08 MB 3.48 KB/s 1.28 GB 28 KB/s 0.019 100.00%
kokrhes No 14.13 MB 69.91 KB/s 0.00 kB 0 B/s Inf. 100.00%
Amalion Yes 13.92 MB 521 B/s 1.28 GB 35.86 KB/s 0.010 100.00%
mama Yes 5.27 MB 0 B/s 118.10 MB 0 B/s 0.044 100.00%
yuppie1977 No 2.66 MB 261 B/s 1.24 GB 186.61 KB/s 0.002 100.00%Now, I was watching, so I know no one jumped on the swarm who got the file already at another site, even though I made it easy for them to do so.
However, I am perplexed by the five I highlighted. the file is 1.28GB, which is what everyone else shows. Not these five, who have all downloaded less than that, usually a lot less. I know that the number would differ if a torrent had multiple files, because not everyone would want everything, but in this case there is only one file.
Is it possible in "torrent world" to simultaneously feed one file from several different trackers? I know it can generally, but I should point out that I changed the torrent slightly. The original video file was contained in a titled folder. I deleted the folder and made my torrent just the video file. I don't know about the pieces breakdown. I do know that the original tracker could not find the file to seed it after I got rid of that uber-folder.
So, to those with more knowledge; were these five jump-seeders, or is something else I don't understand at work.
@ThrobbinRod
Even if we assume these ARE partial jump-seeders, it doesn't seem to have helped them much. Only one of the five has a ratio over 1.0. This can change over time, but I think you were concerned that jump-seeders would suck all the ratio away from "true" downloaders, so you wanted to make a site law against. As you can see, and I have been saying, the people who sucked all the ratio away from the swarm, including from me, were the five "power-seeders" at the top, with ratios from 3.7 to 9.0 and still going strong. Whoever that (anonymous) seeder is at the top, he needs to stop seeding and adjust his client to not hog all the ratio. A site feature that helped this somehow could be useful. -
Hello geekguy,
A part of the answer on the five, is that our records are per session. This means if a leecher stops leeching and later continues, you see only the last stats since he re-started…
Perhaps in this case you were just "lucky" that no high bandwidth member of both sites jumped on ...
Best regards
Uwe -
additionally you have to take into account: the BitTorrent protocoll itself prefers (or is designed to prefer) seeders with high upload speeds. There is nothing we can do against protocol conforming clients.
-
@Uwe:
Hello geekguy,
A part of the answer on the five, is that our records are per session. This means if a leecher stops leeching and later continues, you see only the last stats since he re-started…
Perhaps in this case you were just "lucky" that no high bandwidth member of both sites jumped on ...
Best regards
UweThat makes sense. I knew that for 0.0; I should have mentally extrapolated that was the case.
Mainly I am trying to show ThrobbinRod that jump-seeding isn't a problem like he thinks, because he makes wrong assumptions about where info counts and how things work.
As for the configuration preferences the Mgr mentions: Probably there is nothing to be done explicitly, but as I and others have discussed in the "power seeding" thread, it can impact slower uploaders' ability to share fairly. Maybe its more about educating people to share to a fair ratio, but not an excessive one.
