This repository has been archived by the owner on Nov 15, 2017. It is now read-only.
-
Notifications
You must be signed in to change notification settings - Fork 84
Comparative benchmarks against widely used blockers: Methodology and notes
Raymond Hill edited this page Jan 24, 2014
·
21 revisions
Benchmarks were done using Browser session benchmark.
- The benchmarks are run on Chromium Linux Mint 64-bit. Any Chromium-based browser should be able to run it though.
- Adblock+:
- Allow some non-intrusive advertising was disabled
- Filter lists used:
- HTTPSBOOB and HTTPSBAA/BX:
- Blacklists used:
- [coming...]
-
Fanboy+Easylist-Merged Ultimate List
- HTTPSB uses only the blacklisted domain filters from that list, the rest is entirely ignored.
- Blacklists used:
- Ghostery and Disconnect were set in their respective equivalent of "Block all trackers" mode.
- The latest version of all blockers was used on the day of the benchmark.
- Chrome Settings / Privacy / Content Settings / Plug-ins => "Click to play" was selected.
- Chrome Settings / Privacy / Content Settings / JavaScript => "Allow all sites to run JavaScript" was selected.
- "Adblock+" means Adblock Plus.
- "3rd-party" in the strictest sense, meanings if the domain name of a request differs from the domain of a page URL address, the request is deemed 3rd-party.
- "Cookies" means outbound cookies, i.e. cookies reaching a remote host. These are the cookies which really matter with regard to privacy.
- Keep in mind:
- A user can not add new filters in Ghostery or Disconnect.
- Correct me if I am wrong.
- A user can add new filters in HTTPSB and Adblock+. Where these two differ:
- How easy it is to add/remove filters: Adblock+ = geeky, HTTPSB = easy (just click on a cell in the matrix).
- The granularity of their filters: Adblock+ = coarse- to fine-grained, HTTPSB = hostname/type of request (firewall-like).
- A user can not add new filters in Ghostery or Disconnect.
- HTTPSB, Adblock+ and Disconnect are all GPL.
- I will repeat the benchmark once in a while and add results here in order to show consistentcy of results over time.
- Keep in mind contents of benchmarked web pages will obviously change, so results are meaningful when compared to each other within the same benchmark.