Private vs Public Proxies. Scrapebox Test

I've defended using public proxies a few times on here, and wanted to justify my position by running some tests.

I have access to 10 private proxies from squid proxies on here, and 450 public proxies that took me 30 minutes to harvest and test in SB before running these tests. I'm not trying to compare apples to apples. Individually, one private proxy is better than one public proxy - but that's not what most people have access to.

I tested on two different lists. One with non-unique domains, and one with unique domains. I also tested at 50, 200, and 400 connections.

The private proxies actually did better than I thought they would, and ran really fast for 50 connections or less (so 5 per proxy). Anything above that and their success rate went down considerably.

Public proxies were a lot slower, but were able to maintain high success rates at higher connections. In the end, the highest posts/hr rate came from public proxies at max connections. And, I was using this same proxy list to run one of the other tests at the same time. If I could of run more than 500 connections I think these would have performed even higher.

One other interesting note is that I decided to test the 10 private proxies with just 10 connections. The success rate was extremely high. I would do this again with a high PR list that I wanted to get good results with. However, it was also extremely slow.

So, in conclusion, I think public proxies are still a winner if you have 30 minutes to scan for working ones every day and run them at 400-500 connections, especially if you want to run multiple instances of SB with the same list of proxies.

However, I see two exceptions...
1. You have access to lots of private proxies (like 40+), in which case their speed would eventually overcome their lack of numbers.
2. You have slow internet and can't run 400-500 connections very well anyway, in which case you would get better results at fewer connections with private than public

Here were the direct results:

29k AA blogs, unique url, non-unique domain
10 private proxies
connections success/total total time
10 7790/13126 230 min
35 1586/29521 34 min
50 2096/29521 35 min
200 926/29521 22 min
400 673/29521 15 min

400 public proxies
connections
50 4345/29521 224 min
200 3147/29521 58 min
400 4336/29521 39 min
500 5031/29521 37 min


1.7k AA blogs, unique url, unique domain
10 private proxies
connections
50 138/1677 2.43 min
200 66/1677 1.55 min
400 41/1677 1.15 min

400 public proxies
connections
50 386/1677 16.2 min
200 374/1677 4.26 min
400 384/1677 2.29 min

Definitely agree for harvesting and some add-ons, but not for posting.

soooted said:

Definitely agree for harvesting and some add-ons, but not for posting.

Click to expand...

why?

I have empirical proof to support my position. It may not be air tight - for instance, I ran my test on a VPS and server load probably altered my results at times - but I would like to hear an explanation if you disagree.

efwebs said:

why?.

Click to expand...

Public prox for harvesting so if they get banned you preserve the duration of your private ones?

You didn't check against any blacklists or account for geography. Flawed test and too many connections.

You didn't check against any blacklists or account for geography. Flawed test and too many connections.

You didn't check against any blacklists or account for geography. Flawed test and too many connections.

so whats the concensus on this? I am finally at the point where I need to do some massive posting. Should I buy private proxies? buy shared proxies? use free proxies (lists and such)? harvest proxies with SB? I have 200+ domains that all need extra link love pretty much around the clock (if I could blast one every 2-4 weeks I would be happy). If you have an answer please post some sources too.

Thanks

I just did some massive posting today with public, harvest proxies (didn't bother using them all). I then filtered out the bad ones and duplicates. Then I filtered out the high latency ones. Kept the ones under 800, and blasted away with 30. Used fast posting and had great success. Will never buy private proxies because of this test.

so whats re you basically saying if you want a good success rate of posting comments use -

private proxies:5 connections X the amount of proxies ?
public proxies:500 connextions no matter how many proxies u have ?

also are you referring to the harvest proxy max connection setting? when testing for good proxies or something else..

You don't need 500 proxies, but it doesn't hurt. There's also different sources that you can add to scrapebox's harvester. Just do a careful search on google for a daily list and you'll be fine. Try to use the filter for high latency proxies. First do 2000, and see how much you have left. Then 1000. From here go slightly down and make sure you are left with a good amount of remaining proxies.

also are you referring to the harvest proxy max connection setting? when testing for good proxies or something else..

EisaG said:

also are you referring to the harvest proxy max connection setting? when testing for good proxies or something else..

Click to expand...

i am talking about the clean up filtering after the harvesting

Nice stats, but you're only reporting accepted posts. How do the success rates to actual live links compare? I could imagine that public proxy posts are more likely to be marked as spam.

What annoys me about public proxies is that they are dying so fast. If you are commenting on a large list 100K pages you will notice that if you start with 200-300 working public proxies you will end up at the end of posting with at least 100 short

The thread age is 3 years but a good comparison between private and public proxies.