Hilltop

**John Scott** · 12-02-2003, 12:34 PM

http://www.cs.toronto.edu/%7Egeorgem/hilltop/

Always an interesting read.

webmastery · 12-02-2003, 12:56 PM

Not hard to believe they are using some sort of Expert system now. They appear to be favoring directories as experts on subjects. It also helps explain the significance of outbound linking to authority sites.

**John Scott** · 12-02-2003, 01:49 PM

I find it interesting the way they determine that sites are affiliated.

I, Brian · 12-02-2003, 01:57 PM

The article refers to testing in 1999 - but the high rise of directories in the SERPs is deifnitely suggestive of the early actions mentioned in the article. It would also explain why irrelevant university pages are turning up so high.

Is Florida therefore at least partially cracked, then?

PhilC · 12-02-2003, 02:14 PM

Mel suggested the possibility of the Florida change being Hilltop a while back but, if the -rubbish results are the original derived set of 1000, as I'm confident they are, then the only way I can see the change being due to Hilltop is if Google applies Hilltop when it can come up with a suitably large set of results, and uses the orginally derived set when Hilltop can't manage it. If it's Hilltop, then it appears to be either one or the other, and not a combination when Hilltop can't get a suitably large set.

It could account for why people think that Google keeps a list of searchterms to apply an seo filter to (I've never bought into the list idea). But as to whether or not Florida is partially cracked - I'm afraid not. It could be Hilltop, it could be an seo filter, it could be LocalRank (tho I think not), or it could be something entirely different. It could still be something as simple as a devaluing of same-source links.

**John Scott** · 12-02-2003, 02:22 PM

I don't think that this new algo is so much Hilltop, as it is an SEO filter.

Mel · 12-02-2003, 10:53 PM

It is important to note that the Hilltop theorum has a quite severe limitation - it only performs well on general (as opposed to specific) queries, and for that reason could only be implemented on some searches not all.

Perhaps this could be why some search terms do not change when the -jibbersish term is added?

I, Brian · 12-03-2003, 12:52 AM

I would certainly hope that Hilltop isn't the reason - that would effectively mean that only long-time well established sites therefore have any real link significance. In short, that Google's attempt at "democratic" link voting is decided by a self-interested "upper-class". An oligarchy, in short. And that would be such a female dog to SEO for.

An SEO filter would be much easier to work around, if it exists as suggested.

dpam · 12-03-2003, 11:25 AM

Quote:

An SEO filter would be much easier to work around, if it exists as suggested.

These Google folks clearly just don't get it. How dare they build an SEO filter that can't be easily circumvented with other SEO techniques. Have they no decency (or even a sense of job security)? They're supposed to put up a roadblock, we're supposed to drive around it, they're supposed to move the roadblock, we're supposed to drive around it, etcetera. Searchers keep using Google. We keep making money. This is the system, don't think know that? I'm writing to my Congressman! There must be some way to use the DCMA to sue them for violating my economic good with their digital techniques...

I, Brian · 12-03-2003, 12:13 PM

dpam, you've missed the point. Without SEO, the only pages to consistently rank well for various keywords would be old sh*tty plain text pages from 1997.

There's no democracy in the fact that 7 years ago, when fewer people used the net, single theme sites were more visible. It punishes modern sites for daring to look good and be accessible to a wider and more modern audience.

In short, it's a ridiculous system for a search engine to rank old tatty and now irrelevant pages first than more modern and on-theme ones.

SEO attempts to give those newer sites at least a fighting chance to establish themselves as something worth noting to the contemporary internet audience.

Remember, it's not just commerical sites that seek SEO promotion.

PhilC · 12-03-2003, 12:18 PM

I think that dpam was in tongue-in-cheek mode

I, Brian · 12-03-2003, 12:29 PM

I'm in rant mode.

dpam · 12-04-2003, 10:27 AM

I was goofing on the quote I repeated - that it was too bad there wasn't a way to SEO around the changes that were made to stop SEO. It seemed an ironic, and I'm sure unintentional, way to complain about the update.

Google has an unrealistic view and relationship with SEO, and that is causing all this aggrevation. What gets measured gets done - this is just about a universal truth and certainly applies to Google. They try to circumvent this by not being forthright with 'what' it is they're measuring, but it gets figured out. So then they try to change the rules now and again, and ultimately to come up with measurements that are harder to produce intentionally. I think they're doomed to failure. If 'Florida' is primarily 'Hilltop' (and I think Phil makes a strong case) then Bob Dylan is right once again - 'the game is the same it's just on a different level'.

I'm not smart enough to say what Google should do, and I do think it is their right and duty to get quality results without being 'spammed' or 'tricked' into putting the wrong pages on top. But it seems like if you weed out the top x% of really manipulative tactics, most SEO improves Google's results. I think the process of SEO'ing produces better pages for real people - it makes pages more focused, it forces writing that at least includes the words it's trying to communicate and more often than not probably produces better writing, and increases the inter-linking of sites with accurate labeling of those inter-links. It's not perfect but it tends towards better pages. Moreso than a non-optimized system.

Google threw the baby out with the bath water on this one. I don't think the old results - which were ultimate too manipulated are coming back - but I don't think there is any chance of this system staying without major modifications.

**John Scott** · 12-04-2003, 11:43 AM

Hilltop is oligarchy. It is directly opposed to Google's supposed commitment to a "democracy of the web".

Quote:

PageRank relies on the uniquely democratic nature of the web by using its vast link structure as an indicator of an individual page's value. In essence, Google interprets a link from page A to page B as a vote, by page A, for page B. But, Google looks at more than the sheer volume of votes, or links a page receives; it also analyzes the page that casts the vote. Votes cast by pages that are themselves "important" weigh more heavily and help to make other pages "important."

Important, high-quality sites receive a higher PageRank, which Google remembers each time it conducts a search.

Should be changed to:

Quote:

Hilltop relies on a few biassed opinions. In essence, Google interprets a link from an "expert page", page A, to page B as a vote, by page A, for page B. But, Google ignores the mass of votes existing on the web; it only analyzes the expert pages. Votes cast by pages aren't associated with a .edu are ignored, so please expect an extremely liberal, pro-abortion, anti-gun, pro-gay Internet when you search Google.

janeth · 12-04-2003, 12:37 PM

You guys are way ahead of the rest of the internet they have just started talking about hilltop here now

http://www.webmasterworld.com/forum3/20208-2-15.htm