Is a canonical to itself a link juice leak

[AlanMosley]

Duane Forrester from Bing said that you should not have a canonical pointing back to the same page as it confuses Bingbot,
“A lot of websites have rel=canonicals in place as placeholders within their page code. Its best to leave them blank rather than point them at themselves. Pointing a rel=canonical at the page it is installed in essentially tells us “this page is a copy of itself. Please pass any value from itself to itself.” No need for that.”

He also stated that a canonical is much like a 301 except that it does not physically move the user to the canonical page. This leads me to think that having such a tag may leak link juice. “Please pass any value from itself to itself”
Google has stated that GoogleBot can handle such a tag, but this still does not mean that it is not leaking link juice.

[AlanMosley]

I am a VB man, my algorithm

if url.location.equals(url.relCanonical)  then

leakJuice = 0.15

end if

I say this because of what Duane Forrester said that it assigns its value to itself, all hops leak juice or link juice or it would flow in internal loops, there has to be some decay,

From what he is saying i get the idea it is a link to itself, and would leak.

If you use the Bing SEO Toolkit, it shows you the inlinks, i noticed that there was a link with no link text, i found that this was the canonical tag.

it could be that it assignes value to itself again?

I have removed them from one of my sites and i will see if a get some movement, once i get these ideas in my head i have to find out

[YannickVeys]

I'd say: if url.location==url.relCanonical { leakJuice =0; }

That would be my algorithm for this situation. I don't see any reason why SE's would "punish" you for somethng that could be a mistake. If you rel canonical to a url that is in no way associated to where the tag is placed, I would use the same algorithm.

If the the tag is pointing to a copy of the page somehwere else on that domain I would say: give a minimum link juice leak. If the url is pointing to a different root domain, I would add a little bit more leaking to the algorithm...