What's Wacky with Google?

There are always going to be oddities with any big online service, but this one seems to be persisting. Join the discussion in trying to figure out a pattern. For maybe a week, Google has been returning zero results or "1-1 of about xxx,000" for common searches. One-word searches seem unaffected, but there are certain two-word combinations of common words like candle truck or speaker bracelet. Reversing the order can affect searches too: motorcycle candles vs. candles motorcycle. The strange thing is that usually the 1 or 2 results found are to commerce sites. Read the Search Basics, compare your notes to GoogleWhack's, have fun looking for patterns, but remember that Google always returns slightly different results for different IP numbers.

(Update: 13:56 GMT by J : When I first posted this story it said the problems have been occurring "for several weeks at least" -- but it seems to be more like one week.)

Man!

I am so glad someone else noticed this!!! I've been so pissed I haven't been able to get any speaker bracelets recently. God google... forcing me to use other search engines to get my fix.

Re:Man!

And I've been trying to get rid of several pallets full of candles I've got sitting around the house, but haven't been able to find a truck suitable for the job. I need to get these out into the market, since I went to all that trouble to install RFID tags on each one...

Re:Something I've noticed recently...

Thats because "vb.net" is a URL, and google treats it like one. It actually returns exactly one result, a link to vb.net [vb.net], as it should.

If you're looking for the product "VB.NET", you need to search for it as a term.

Re:Something I've noticed recently...

...is that a search for VB.NET does not return any results either unless you perform an "Exact Phrase" search.

For ordinary searches, punctuation marks like "." are treated as spaces, which mean logical ANDs. And some words (in this case "vb" and "net") are ignored as being too common. If you search for "vb.net", which I suppose is what you get from an "exact phrase", you find "vb" followed by a space or punctuation and then "net".

Google tries to be intuitive, which means guessing what most people would expect, which of course means that sometimes you're surprised.

It's obvious

SkyNet is becoming self-aware.

Re:It's obvious

Yeah, and it has already aquired the targetting mechanism of Half-Life 2... 8)

Deja vu?

It's just a glitch in The Matrix, of course.

Re:It still can't do phrase searches

At the risk of making you look bad, for phrase searches you have to put the phrase in quotes.

For example, I searched for "to be or not to be" phrase origin [google.com], and got what I consider to be useful results.

YMMV, of course.

Xentax

"to be or not to be"...

should produce about 50% error rate or we are really in trouble ;-)

Re:"to be or not to be"...

2*b || !(2*b) is actually a tautology :P

ducks :P

Re:It still can't do phrase searches

Blockquoth the poster:

Is 100% accurate matching results to a phrase search too much to ask for a search engine?

Um, yeah. Actually, I don't know what you're talking about. Entering the phrase "to be or not to be" -- with quotes, so as to indicate you want the phrase, not just the collection of words -- yielded the first two pages of results all having that phrase. Not all of them were for pages on Shakespeare, but then again, that phrase is now deeply buried in the common memespace. If you make the search phrase

"to be or not to be" Shakespeare

you do indeed get results with the phrase and exclusively referring to Shakespeare. Oh, I get it. You don't like the idea you need to actually construct a reasonable search phrase. You're mad that Google isn't, I don't know, telepathic. Your best bet is the SFWIWNFWIS search engine -- search for what I want, not for what I say.

Re:It still can't do phrase searches

Google doesn't do simplistic phrase matching. If it did, it'd be the same (and as useless) as altavista. Google does relevancy searches. tobeornottobe.com is relevent to a search for "to be or not to be".

Hello? You're kidding, right?

What a cockamamy way to run a search engine.

You are kidding, right? There's a reason that Google is by far the most popular search engine on the web, and it's got a lot to do with the "cockamamy" way it's run.

Perhaps you prefer the good old days when you'd have to check half a dozen search engines and trawl through countless useless links until you found something that was useful.

There are a handful of websites that should be in everyone's bookmarks. Top of the list is Google. Nuff said.

Oh, and as several people will have mentioned by now, and as Google's FAQ surely does, putting your search parameter in quotes will give you exact phrase results. This is pretty standard amongst all search engines, so it's amazing that you don't know this already.

Either you're new to the web and search engines in general or you haven't got a clue how to use one. Regardless, if you're going to comment on how "cockamamy" Google is, you should at least have an idea of how to use it first.

Candle Truck? Speaker bracelet?!?!

What possesses someone to try such weird random words in google. Thats the real trick...google wrote an engine to amuse the crazy users.

Corporate entity

That's why you can't trust google for anything critical. You are at their mercy, and if they choose to do biased, or screwed up searches, you either don't know, or can't do anything about it...

I propose an opensource web based search engine... No more weirdness, no more screwups, no more censorship!

Re:Corporate entity

Slashdot (tm). Where reality is (-1, Flamebait).

Seriously, you know what make google so great? Part of it's the interface. Part of it's the software. But most of it is the company. The clout to afford enough bandwidth to spider the earth on a routine basis. The cash to maintain thousands of servers and a complicated database with which to serve not only their engine, but a CACHE of pretty much everything they index.

No open source project will ever have the ability to do these things. Because the people who are good enough salesmen to get the revenue needed to do what google does won't want to dillute their position by allowing any hacker with a gimpbox to run the same engine. And the people who are good enough open source software designers to write an engine like google wouldn't want some ad guy treating their work like it was inktomi. You can't run a search engine without money, and you can't run an OSS project like a truly commercial enterprise.

At the end of the day, distributed software doesn't lend itself well to large, FAST, searchable databases. And if this is -1, Flamebait, I guess you may flame away.

Google Zeitgeist

I am sure the next Google Zeitgeist [google.com] will show numerous searches for candle truck or speaker bracelet in October 2003. And nobody at Google will have an explanation for this ;-)

Re:Google Zeitgeist

quiet fool! .. you've uncovered the true plot behind this slashdot posting.

No longer will we /. sites to take them down .. we will effect data mining for common searches on the internet.

Long live the Speaker Bracelet

The same words in quotes show more hits ...

Check out this [google.com] - all 25 hits on the quoted words "candle truck" should be showing up in the non-quoted search ...

One of the hits

Not wanting to kill anybody, we wait until the last two guys wander up to the candle truck.

I prefer not to even click on that one, and just speculate.

maybe

Maybe it has something to do with the counter that was meantioned in a slashdot post earlyer today?

Re:maybe

Are you suggesting they are trying to count the candle trucks?

groups/deja is also acting up

for a few weeks, when I do a search on google groups, it'll come back with the results just fine - but when I click on the View Thread on a result, it tells me it can't display the thread and gives me a link to view that individual message. Then once that message comes up, I click on View Thread on that message, and up pops the whole thread, like it should have before.

Perhaps being on the top is getting to their CPU's :)

Re:groups/deja is also acting up

I've been getting oddities there as well, although it usually just doesn't show anything. A reload usually shows the thread correctly. I wrote it down to busy servers or some other transient fault; perhaps there's a larger fault somewhere in Google? I certainly hope not.

Another oddity has been that threads have been stated as having "1 post", but viewing the thread shows a larger thread.

Re:groups/deja is also acting up

Google are aware of this problem and are working on it. I know cause I wrote to them with some example URIs and they replied they are working on some known issues with their servers.

Re:What's wacky with Slashdot?

Since you asked :)

No, stories don't have to move through the cluster, and there's no concurrency bug. We have a front-end cluster of webheads but they all read from the same DBs. The only "moving through" is from our main DB to our replicated slave reader DBs, but they are typically only 0 to 1 seconds behind reality, so that's not an issue.

In this case, the problem was that Hemos and I were both editing the story at the same time. He added an icon and posted it at 9:36 EDT live, then I tweaked the text and posted it at 9:38 which was about 40 seconds in the future, then around 9:39 I went back and edited its time back to 9:36... so there were a few seconds there where the story went from front-page to subscriber-only and back.

The Slash backend is obviously too powerful for idiots like us :)

Google Whackiness

Has anyone else noticed that the "spam" sort of sites that are nothing but link farms and Gator popups are getting much better at finding their way into Google's rankings? I switched to Google back in the day after search engines like altavista became overrun with such sites. Now I've noticed that they occasionally creep into their rankings...I guess entropy is the way of the universe after all.

Re:Google Whackiness

I too am experiencing this more and more during the past few weeks (months?)..

For example when searching for visual basic decompiler [google.com] the second to fourth results are 'spam sites'.

I always report this kind of crap via the "Dissatisfied with your search results?"-link, but apparently nothing is done against this sites, which are getting more and more annoying.

Time to switch? [teoma.com]

Gator and Zuvio

I have run into some bizarre results lately. Recently I was trying to figure out what the NT 4 process "ESSERVER.EXE [google.com]" did, and google's top search result sent me to a page at (DON'T GO HERE!!!)MamuFilms.com [mamufilms.com] which actually redirects to "Armbender.com", a site that won't show you any pages unless you install "Page Access", actually Zuvio nastyware.

Here's Googles somewhat hilarious cache [216.239.37.104] of the Mamufilms.com page. The page includes links for everything from "Peter Paul and Mary mp3" to "preteen bra images". The text is vaguely reminiscent of actual gramatical English. Here's one sentence:

And With Unknown virtual gifts Already baby food coupons to Information Installed The 2000, with Himself, to other tips, tricks, and tweaks The Issue De Processes services.exe.

Re:My results for "candle truck" (completely genui

scented candle truck accessories

Wow, talk about a niche market...

Another thing - what triggers the calculator?

I realized the other day that although searching for 13 - 867 - 5309 [google.com] causes google to go into calculator mode, searching for 123 - 867 - 5309 [google.com] does not cause google to use calculator mode.

All sorts of odd things will both pull up an answer from google's calculator and also do a search - for example, searching for avogadros number [google.com] or hbar [google.com].

So why do searches that might fit US telephone conventions not trigger calculator? Is it because some design decision makes it impossible to trigger both calculator and their phone lookup service. (Yes kids, google is a reverse phone directory, albeit with old data)

What's wrong with this picture?

"q=site:www.google.com google [google.com]" - (third result)

This is what I'm seeing...
http://www.sminkybang.com/google.png [sminkybang.com]

General idea:

google uses tons of DB entries to cross-index pages. I wonder if there's some simple hash-tables per page that it uses internally to speed things up that makes assumptions, and doesn't resolve collisions.

So you can search for one thing, and conceivably the checksum/hashes for each term match those of another page that has nothing to do with it, and it's returned as a relevant match by accident.

This might explain a lot of result sillyness.

On Google buying Kaltix

By the way, for info on Google's purchase of the search engine Kaltix, check this controversial Register piece [theregister.co.uk] by Andrew Orlowski. It contains the highly suspect, matter-of-fact comment that "PageRank is now widely acknowledged to be broken," but if you take the PageRank speculation with a grain of salt it's an interesting read.

Canuck Ok

For any who are interested, Google.ca is behaving correctly. All search results listed (that I've tried so far) from googlewack.com are working properly and returning 1-1 of 1, or displaying as they should.
I wish I could compare to google.com, but for the past year or so, google.com automatically forwards all canadian IP's to google.ca

Re:Canuck Ok

Put

216.239.37.99 www.google.com

In your hosts file to force it to resolve to the US google, or just type that in your browser.

Alternately you can search google for the other googles and connect to them through google, for google japan, google australia, or google canada for example - or you can just hit the go to google.com link at the bottom of the google.ca page which links to http://www.google.com/ncr which I guess disables the country recognition and could be used as a bookmark as an alternative to modifying the hosts file.

What's wacky with slashdot?

Does anybody else see the story change? I'm getting two different versions if I reload. One with the additional lines:

"The order of words matters also, with motorcycle candle revealing different results to candle motorcycle."

"Read the Search Basics, compare your notes to GoogleWhack's"

and one without.

Complete text of the two versions are:

"There are always going to be oddities with any big online service, but this one seems to be persisting. Join the discussion in trying to figure out a pattern. For several weeks at least, Google has been returning zero results or "1-1 of about xxx,000" for common searches. One-word searches seem unaffected, but certain two-word combinations of common words like candle truck or speaker bracelet are affected. The strange thing is that usually the 1 or 2 results found are to commerce sites. Have fun looking for patterns but remember that Google always returns slightly different results for different IP numbers."

and

"There are always going to be oddities with any big online service, but this one seems to be persisting. Join the discussion in trying to figure out a pattern. For several weeks at least, Google has been returning zero results or "1-1 of about xxx,000" for common searches. One-word searches seem unaffected, but there are certain two-word combinations of common words like candle truck or speaker bracelet. Reversing the order can affect searches too: motorcycle candles vs. candles motorcycle. The strange thing is that usually the 1 or 2 results found are to commerce sites. Read the Search Basics, compare your notes to GoogleWhack's, have fun looking for patterns, but remember that Google always returns slightly different results for different IP numbers."

Strange.

COMMON searches?

Ok, now I'm a guy who deals with audio equipment on a regular basis. This, of course, includes speakers. I have never, ever, heard of a speaker bracelet, and can't imagine why one would search for it.

Now this isn't to say that these people havn't perhaps discovered an interesting bug in Google, but trying to play it as a conspiracy for "common" search terms is bullshit. The terms listed are things that no normal person would EVER search for. Hell, they are terms that even someone involved with one of the terms would never search for. Bracelets have nothing to do with speakers. If Google was truly trying to push advertisers, well, they'd be doing a shitty job of it since only geeks with too much time on their hands would discover such things.

Give it a rest, the world is not out to get you. It's either a bug, or Google having some fun (something they are known to do). They are certinaly not trying to pimp a certian manufacturer of speaker bracelets, since such a thing is something that noone would know about, care about or want to own.

For regular searches, Google continues to work great.

Simple.

It's broke. Just put a sign on it and someone call the super.

speakerbracelet.com

Mwahhahahah!

1. Register speakerbracelet.com

2. Be the top 1 of 2 search results on google.

3. ????

4. Profit!

The real time search monitor

I've read that there's a real time search monitor in the lobby of Google's HQ. The nastiest words are removed, but other than that you can se exactly what people are searching for.
They have to be pretty confused right now, when thousands of searches for speaker bracelets, motorcycle candles and candle trucks show up on the display!

stone dog quote

searching google for stone dog quote returns no results. Also try stone cat quote or changing the order of the words for weird results. Queries on alltheweb or altavista return numerous results, as expected. This has been reported in threads in alt.usage.english, rec.puzzles and (of all places) alt.fan.tolkien.

speaker bracelet two

Results 1 - 10 of about 27,300

Weird. Very weird. Adding another word to a search should narrow down the result set, not widen it.

Try [google.com] it.

Re:speaker bracelet two

gm candle truck: Results 1 - 10 of about 12,100
fiat candle truck: Results 1 - 10 of about 5,200
audi candle truck: Results 1 - 10 of about 7,090
chrysler candle truck: Results 1 - 10 of about 18,400
ferrari candle truck: Results 1 - 10 of about 9,810
ford candle truck: Your search - ford candle truck - did not match any documents.

Looks like it's about time ford got on the candle truck bandwagon.

Is it a glitch?

.....in the Matrix? Have we found it?

Maybe they're tweaking

Obviously, Google has to do a lot of acrobatics to keep its service as fast as possible. One of the things it does is distributing its database over a lot of servers. There is no way that they can dynamically sift through hundreds of millions of pages for each common word, so they obviously just look at the top pages for each word. Which pages are top is probably determined by pagerank or something similar.

When you do this, there is no guarantee that you will get hits for every single combination of words out there. However, it may very well be possible to calculate the probability of relevant results not showing up and using this measure to make a more or less optimal trade-off between response time and user satisfaction.

When you start tweaking this trade-off, certain queries are bound to get screwed up. It probably takes them some time to notice this behavior, gather statistics and re-tweak their formula.

Another thing that crossed my mind recently is that they might be using precooked phrases or word collocations instead of single words. This makes sense since they use an implicit AND operator, it improves statistics and words are often strongly correlated anyway so your vocabulary probably wouldn't swell as much as you'd expect.

Mind you, this is pure speculation. I don't have any intimate knowledge about Google's inner workings.

Pending purchase?

Maybe they are doing testing to ensure the service will be in the usual Microsoft standards in case they are bought.

Another example

A few days ago I searched for "kazaa lite [google.com]" on Google and found that no results are censored! The main KaZaA Lite page was the 1st result. That was only temporarily, of course, because right now the search is still censored.

Who cares about searching for that.

I used to use motorcycle candles but later found that an electric headlight is much brighter and doesn't blow out when I move.

Candle Truck?

Man, no wonder... You need to turn Safe Search OFF when you look up nasty stuff like that.

Not a week . . .

(Update: 13:56 GMT by J: When I first posted this story it said the problems have been occurring "for several weeks at least" -- but it seems to be more like one week.)

Actually, I've been seeing this problem occasionally for over a year. It just seems that larger numbers of search terms trigger it now.

Of course, I can't remember any of the search terms that have triggered it in the past--I've just learned to change my terms slightly to get around the problem.

Dee

Maybe it isn't a problem!

Google searches use unique and proprietary algorithms to find the most useful information for the search terms. We all know this, it is their "page rank" system. But perhaps the page rank system is driven by more modifiers than we are aware of. For instance, In Minnesota, Twins and Vikings mean a couple of sports teams, in Norway, they probably mean something entirely different so perhaps "Page Rank" does some regionalization. In the same vein, it may be possible that if I refine my search from Minnesota by adding the word "Gopher" to the Twins and Vikings, I may get more, rather than fewer results while perhaps in Norway I'd get no results!

In addition to possibly regionalizing searches, perhaps Google's servers are not updated with the latest code at the same time. Maybe the code is distributed over time to servers so that if a problem were discovered it could be more easily rolled back. It is possible that the load balancing on these servers uses some component of the IP address or somehow regionalizes the incomming requests so that it is likely that the same user usually gets to server A but sometimes goes to server B while their co-surfer neighbor usually goes to server B but sometimes goes to server C. Meanwhile, a couple of states away, another user usually connects to server W but sometimes connects to server X. This could explain why they usually but not always get the same results but someone else gets different results.

The *REAL* answer

The pigeons [google.com] are getting tired.

Real information

I spoke with a friend who helps maintain the google engine. She said that they were running into some problems with a "cleaning agent." Because of all the sites taking advantage of the word revelancy, there are useless sites that simply have a list of words or phrases. It's been posted before that there are many pages designed for GATOR/GAIN spreading or other spyware/adware. She quoted the percentage of junk pages being at 35% to 40%. The cleaning agent was supposed to run through its own searches and check for junk and keep a log.


She didn't say if the problem was that the cleaning agent was clogging searches or if any logged junk pages had been blocked. If so maybe the agent is flawed. In any case, they've stopped using it for the time being.

Strange counts for five weeks now

The counts have been broken for the last five weeks. A count for the word "the" produced fairly consistent results until then of about 3.4 billion. Then it shifted five weeks ago to 5.2 billion. Lately it has been under 2 billion. Now it's just over 2 billion.

Webmasters who have various directories and know exactly how many pages are in each directory, began noticing five weeks ago that Google was reporting approximately twice the number of pages in each directory than have ever existed in that directory. Prior to five weeks ago, Google used to be fairly close to the actual number (assuming that you get a full crawl).

GoogleWatch speculates on the reason why Google has been behaving strangely ever since it stopped doing the traditional deep crawl once per month. The last standard deep crawl was in April but it wasn't used -- Google threw out this data (by their own admission) and reverted to earlier data. The speculative piece [google-watch.org] was written last June.

Since it was written, Google has started showing "supplemental results" on many searches. It looks like they are running a parallel index. Why would they do this? All the problems Google has been having, along with the supplemental index, seem to support GoogleWatch's theory.

Strange results for duplicate search terms

Duplicating search terms has an interesting result:

candle truck
1-1 of about 101,000

candle candle truck truck
1-1 of about 82,200

candle candle candle truck truck truck
1-1 of about 73,700

candle candle candle candle truck truck truck truck
1-1 of about 68,600

Another interesting one is

candle candle truck
1-2 of about 89,200