HarborFront

Level 53
Verified
Content Creator
This is not true, let me explain my point (start of rant ;) , not against Handsome recluse, but against "more is less, benefits of large block-list perception" in general)

There are nearly three billion websites, but in the Western world 2 (two!) companies serve half of the ads (Google and Facebook). The top 100 advertising networks have a market share of around 80%, the top 200 around 90%. Lists with 3000 to 4000 rules (Peter Low's, Disconnect, Ghostery) block 95% of the advertisements. With 40.000 rules (Adguard optimzed) you might block 97% and with 1.5 million 98%.

It is not about blocking individual ads on websites, but blocking third-party access of advertising networks on these websites. Visitors are tracked on websites. Those visitors are served (interest based) ads not by the website they are visiting, but by the advertising network which pushes ads to the website you are visiting. It costs money to build an ad-serving infrastructure. For displaying an advertisement (and more important a click through) websites get paid in cents, so you need a lot of traffic to earn money from it.

These two factors, the money involved to build an advertising network and the low payment per advertising display, is the reason that there are a limited number of advertising networks. In every market in the capitalistic world the market share is divided by a limited number of players. As a rule of thumb the higher the price of infrastructure or production of a product (e.g. 5G network providers or car manufacturers), the lower the number of market players. Same applies on market maturity, the older the market the lower the number of players involved (e.g. vendors of branded diary products).

It is the same with payment services (see post: Q&A - List of digital payment services per country ) In the Netherlands less than 10 players serve maybe 20 million websites. Simply because it takes money to build such an infrastructure. When you want to prevent your family members from buying online, you don't need a block list of 20.000.0000 rules, just block the 10 digital payment services servicing the Dutch websites. For the Netherlands a block-list of 40 third-party block rules (listing banks and payment services operating in the Netherlands) will probably be sufficient to prevent all your family members buying on-line on 95% of the 20 million Dutch websites. Hunting the last 5% down, would cost a lot more effort. It is the same with advertisements, the effort to catch the last 10 to 5 percent of the long tail is tremendous.

Normal market mechanism apply to the digital market as well. In most markets 20% of the players account for 80% of the turnover. The remaining 20% is divided between a lot of smaller players. Because of the low prevalence of those small players, those smaller ad-networks are not blocked by ad-blockers and serve a specific niche or language market. Only when you visit a lot of niche websites, large block list might be beneficial. But the odds such a niche player is included in a large block-list are still low (a block-list of 300.0000 website specific rules only covers 0,001 percent of the websites in the world). The reverse is true: people with large block-list over-optimize: they arm themselves against websites they never visit and not-curated dead ad-links (curation is a major problem of community based block-list).

My take on less is more versus more is less:
  1. W3Tech top 200 list is an example of a bare minimum (the bottom range and minimalist take of/on less is more)
  2. Peter Low's, Ghostery and Disconnect (with < 4000 rules) block without breaking (reason both Firefox and Edge-Chromium use Disconnect), These list are the representatives of the main stream (less is more) approach.
  3. Adguard's optimized filters (less than 40.000 rules) is the optimum where less is more and more is less meet. Adguard gets feedback about the (originally Easylist rules) which are really used, making Adguard's optimized filters currently the best in the market (highes blockrate with lowest functional breakage). The automated feedback mechanism skips/omits websites with low traffic and dead ad-links. A simular optimum can be achieved with uBlockOrigin using uBO's own filter plus Steven Black's hostlist (but you have to know how to add that) and everything else disabled.
  4. uBlock's with its own uB0-filters and only Peter Low's and Easylist ads plus trackers is probably the optimum for a more is less approach.
  5. uBlock's default without the malware protection is probably the maximum useful of more is less approach. Adding more is futile.

Hope people understand why less is more is 95% effective as more is less and chasing the last 5% is futile (end of rant :) )

P.S. with a block-list of under 250 rules, I never run into anti-ad-block walls, another advantage of a small block-list!
I see you quoted W3Tech list (or rather the list by @Windows_Security). According to the link below both the lists are updated once a year only as stated by @Windows_Security


I don't feel comfortable for a list which updates only once a year
 

Lenny_Fox

Level 11
I don't feel comfortable for a list which updates only once a year
Ahhh my rant was in vain :) , reason for low update frequency
1. The W3Tech survey on which W_S based is held only once a year. As far as I understood W_S just imports the results.
2. The players in adserving networks don't change much (BECAUSE IT COSTS A LOT OF MONEY TO SET UP SUCH A NETWORK !!!)

This list does not hunt down the websites with ads, but the advertisement networks serving these ads to websites (you visit)

For those mega lists fans, It is dry January, why don't you detox on block-lists and run solely this tiny block-list for a few days, you will be surprised.
 
Last edited:

SeriousHoax

Level 29
Verified
Malware Tester
Hi @Lenny_Linux shade some of your adblocking knowledge here. According to Steven Black, 0.0.0.0 is faster than 127.0.0.1 but Peter Lowe's list added by default in uBlock Origin uses 127.0.0.1. I searched for a 0.0.0.0 format for Peter's list but couldn't found one. But I found one based on Adblockplus rules: Blocklist of hostnames and domains for blocking ads, trackers and others (format: adblockplus - for use with Adblock Plus)
Now by using this AdblockPlus one or a 0.0.0.0 one over 127.0.0.1 would really make any difference or not?
Btw, I also use SimpleDnsCrypt which is set to use cloudflare but locally has set 127.0.0.1 as my dns server. Would this make any difference in terms of speed as well for the Peter Lowe's host in uBlock Origin?
 
Last edited:

Lenny_Fox

Level 11
@SeriousHoax

I don't think the block-list format makes a lot of difference. I think uBlock interpretate's host file blocklists as very simpe ABP block-lists. So I don't think it matters whether you use host file or ABP format in uBlock. My guess is that there is a performance difference when comparing uBlock with uMatrix in the benefit of uMatrix,.

The dynamic blocking engines of uBlock and uMatrix are more or less the same with uMatrix having more granular control, so in performance they should be equal. When you compare the static engine of uBlock with uMatrix the uMatrx versions is much simpler. So uMatrix should use less CPU when chewing on the same blocklists as uBlock.

When I look at the Javascript on Github (for as much as I understand, only reading JS code) the JS-code of uBlock related to static filtering is at least 3 times larger than the one from uMatrix. Due to more complex ABP syntax, the data structures to process are also more complex so uMatrix is probably faster than uBlock.
 

HarborFront

Level 53
Verified
Content Creator
And 3 gifts - tests for you:
thepcspy.com/blockadblock/: https://thepcspy.com/blockadblock/
immediate.co.uk: Ad Blocker Interference Detected: Immediate Media Co – the special interest content and platform company
playhydrax.com: https://playhydrax.com/
Anti-adblockers work based on lists. If the sites are on the lists they'll be prevented from seeing the 'AdBlock Detected' message otherwise the aforementioned message will show up

To me this is not a critical issue. If I want to surf the site then either I whitelist the site or turn off my adblocker otherwise I'll give the site a pass
 

SeriousHoax

Level 29
Verified
Malware Tester
@SeriousHoax

I don't think the block-list format makes a lot of difference. I think uBlock interpretate's host file blocklists as very simpe ABP block-lists. So I don't think it matters whether you use host file or ABP format in uBlock. My guess is that there is a performance difference when comparing uBlock with uMatrix in the benefit of uMatrix,.

The dynamic blocking engines of uBlock and uMatrix are more or less the same with uMatrix having more granular control, so in performance they should be equal. When you compare the static engine of uBlock with uMatrix the uMatrx versions is much simpler. So uMatrix should use less CPU when chewing on the same blocklists as uBlock.

When I look at the Javascript on Github (for as much as I understand, only reading JS code) the JS-code of uBlock related to static filtering is at least 3 times larger than the one from uMatrix. Due to more complex ABP syntax, the data structures to process are also more complex so uMatrix is probably faster than uBlock.
Thanks.
Hmm it makes sense since uMatrix is host based only and also lacks cosmetic filtering.
Anyway, found 0.0.0.0 format of the Peter Lowe's list if anyone's interested: https://pgl.yoyo.org/adservers/serverlist.php?hostformat=hosts&mimetype=plaintext&useip=0.0.0.0
 

SeriousHoax

Level 29
Verified
Malware Tester
Here's mine. I don't use much host based list except the default Peter Lowe's list and abuse.ch URLhaus Online Malicious URL Blocklist because I've a huge host list imported in SimpleDnsCrypt containing over 680K domains which I update twice a week.
View attachment 232289View attachment 232290
Replaced SimpleDnsCrypt with Adguard Home. Everything is automatic now 👌
You guys may try it as well for system wide host based blocking and reduce the number of host based list in uBlock Origin.
 
Last edited:

Handsome Recluse

Level 23
Verified
more is less, benefits of large block-list perception
My point still stands. What benchmarks are you aiming for? Less RAM/CPU usage? More relevant hours in a week saved? Or just shorter lists?

Replaced SimpleDnsCrypt with Adguard Home. Everything is automatic now 👌
You guys may try it as well for system wide host based blocking and reduce the number of host based list in uBlock Origin.
On a seperate device or your personal PC?
 

SeriousHoax

Level 29
Verified
Malware Tester
On a seperate device or your personal PC?
On my PC only. I have NextDns set on my smartphone and on router for my family members with many filters enabled including porn blocking. But NextDns is slower than Cloudflare for me so on my PC I use cloudflare.
With SimleDnsCrypt I had to update my host manually but with Adguard Home it's automatic so less effort to maintain and also has fantastic query logs which is very helpful. Of course it supports encrypted DNS and unlike SimpleDnsCrypt it supports DoT along with DoH and DnsCrypt but it's not mandatory to use encrypted dns, you can use without encryption.
 
Last edited:

HarborFront

Level 53
Verified
Content Creator
You can install it on Linux, Windows, etc. It's not a stretch for him to think "Oh, Adguard Home supports encrypted DNS. Let me install it on my PC instead of SimpleDNSCrypt"
Adguard Home looks like not supported on Windows

Quote

How to set it up

Currently, you can install AdGuard Home on MacOS, Linux (x32 or x64), and Raspberry Pi. There is no need to go into technical details — we have the installation process described in the dedicated GitHub repository. Also, detailed guides on installing AG DNS to VPS and Raspberry Pi are available in our Wiki.

Unquote

 

Lenny_Fox

Level 11
:confused: Help me understand: Is the tip for uBlockOrigin/NanoDefender uses is to de-install both and install Adguard Home or is this a (competitor) thread topic take over or is it an example of security forum new-program-to-play-with frenzy? :)
 
Top