25 October, 2005

Blogelo: I Call Shenanigans

Blogelo is the Hula Hoop of Today.

Yeah, so what....I wasn't even listed until I added myself. Not that I suffer from an overinflated sense of self-importance or anything. Unlike some people I didn't spend all day trying to boost my ranking. Primarily because I didn't, and still dont, understand how.

Now, I get that the whole deal is based on the Elo system for rating comparative players. I first heard of Elo in relation to Go several years ago, and now it's everywhere in relation to chess.

It strikes me as odd to try to forcibly apply Elo to blogging because it's a Game Theory measurement. It involves comparative rankings based on a system of wins and losses. Now, I've had some bad posts in my day, up to and most probably including this one. However, posting is not a zero-sum game. Okay. Sometimes it's a zero-all game. But we're not talking about Daily Kos right now.

Right now we are talking about how I feel like an overgrown toddler easily distracted by sparkly colors on shiny objects. "Look, mommy! There's my blog. It's ranked number 27." I have no idea how this ranking happened...external links? Sitemeter hits? number of people who generally think I'm an jerk? Cause I gotta tell you, if we're going with the Jerk Index I woulda thought Ida been higher up.

Please, Some Geek In Tennessee, please explain to us bottom feeders (okay, just me....) how on earth this works and how in the world we can all gang up and unseat the Vandy Alum Champ.

UPDATE

So he explains it all right here in the blog section. Good to know that I figured out how to read this late in the game. Still not seeing how a zero-sum rating system can be strictly applied to blogging, though. But I know in my heart I'm as good as, say, Slashdot. hahahahahahaha!!!!!

15 Comments:

At 1:27 AM, October 25, 2005, Blogger Sharon Cobb said...

I don't understand most of it. Another list on it showed me in the top 10, whie another showed me at 70. I have no idea what it means...either one. More importantly, does it matter?
It's too confusing and unimportant.

 
At 1:31 AM, October 25, 2005, Blogger Kat Coble said...

I think I get it now, because I was just over there watching it.

You pick two sites and click on the one you like better. That drives it up in the list of rankings.

So....what happens when we all rate our own sites as "best"?

Right now it looks like the lead guy is going back through and drilling down Kleinheider's ballot-box stuffing.

If you click on "spy" you can see who exactly is voting for what and how.

 
At 3:03 AM, October 25, 2005, Blogger Sharon Cobb said...

I see. And this is a good way to spend our time because______?

If I could write as well as you, I'd never stop writing in the first person or in the narrative form.

I'll make you a deal. I'll vote for you on this thingie so you don't have to take time away from your writing.

 
At 8:59 AM, October 25, 2005, Blogger Patrick said...

I'm bigger than Flickr! (as of 9:00 this morning.)

 
At 9:04 AM, October 25, 2005, Blogger Patrick said...

See? (for posterity)

 
At 10:58 AM, October 25, 2005, Blogger Kat Coble said...

Someone is massaging his numbers, methinks.

Not that I haven't been after figuring out how....

 
At 10:59 AM, October 25, 2005, Blogger Casey said...

Well if it ends up just being a tool for self promotion then my little experiment is a failure. Looks like that the way it's heading however so I may just have to live with the limitations of del.icio.us :(

 
At 11:03 AM, October 25, 2005, Blogger Kat Coble said...

I think it would work if you were the sole voter. But giving everyone a chance to vote across all categories is where it falls short.

Taking it back to the chess realm, it would be like letting individual chess masters rank themselves against Go players, popular MMMORPGers and shoe salesmen.

The categories are too broad and the voters too many to narrow down the stream into ELO mathematics in my opinion. But it is a good idea in theory, if you can figure out a way to narrow down the voters and the categories.

 
At 11:07 AM, October 25, 2005, Blogger Casey said...

Let me add, I was hoping more folks would end up using it as I am. Pretty naive, I know. I don't vote my own sites up or down. I know that SGnTN and blogelo are the greatest thing since sliced bread of course, doesn't matter what anyone else thinks.

Anyway, I just compare those that I know a little something about (for the most part) and hoped that everyones combined opinions would come together to create a useful resource for people looking for information about a particular topic.

 
At 11:14 AM, October 25, 2005, Blogger Kat Coble said...

I voted my own site up only when I honestly preferred it to the site I was voting against. Even though it IS shenanigans to compare my personal blog to Rotten Tomatoes, for instance.

Other cases I called a draw. Like with Lileks. Egotistical, since he's the better writer, but that is my own pride getting in the way.

Seriously, though, in Thunderdome, most people are going to massage their numbers massively, which invalidates the experiment.

The part I like best about the experiment is being able to compare blogs/sites by tags. That is very very handy.

 
At 11:14 AM, October 25, 2005, Blogger Casey said...

Well I used ELO as a starting point. If you're familiar with the formula, then you know there is a multplier called K that bascially describes the weighting of the match. I adjust the K factor based on a number of factors such as how many matches the winner has been in, and the number of tags the voter has applied to both contestants, etc etc.

If I stumble across a better rating system maybe I'll plug that in. What yout typically see on the net is a "rate this from 1-10" type thing and it's completly arbitrary and useless as far as I'm concerned.

Oh an btw, the ratings for the various categories (tags) are not applied unless both sites being compared are in the same category. A site tagged dogs compared to a site tagged cats will not have it's rating adjusted in either category. Just the overall rating.

 
At 11:18 AM, October 25, 2005, Blogger Kat Coble said...

Oh an btw, the ratings for the various categories (tags) are not applied unless both sites being compared are in the same category.

Gotcha. That does make cleaner comparisons.

Is there any way to set it up so that people can vote on everything except their own properties, as it were? Cause if so, that would solve the number massaging problem to a great deal.

The problem with adjusting the K factor for the number of matches is what is banging up against the personal voting, because it allows the multivoter an exponential gap.

 
At 11:30 AM, October 25, 2005, Blogger Casey said...

OK real quick before that lunch thing...


If I made it where people couldn't vote on their own sites then I'd have to figure some way to prevent them from setting up another user and using that one to promote their site. If I set some arbitrary limit for the number of votes.. well that would suck. One possibilty is that each succesive vote for a particular site within a certain amount of time adjusts the K down.

As for K / number of matches. If you imagine this site with lots and lots of people using it, then there would be plenty of votes for sites like /., digg, wikipedia, etc. Someone looking to promote themselves by comparing themselves to these sites would get rewarded far less for a win that those sites that were already popular. Trouble is, AC has beaten the crowd, but if blogelo works his votes will start to get lost in the noise and things will sort themselves out... um.. maybe :)

OK I'm hungry and need some Thai food!

 
At 12:44 PM, October 25, 2005, Blogger Casey said...

This may be an after effect from the Thai food but...

Here I was all excited after the buzz around blogelo yesterday.

It may be that it's just to complicated to appeal to the everyday user which is just the sort of person I need for it to appeal to if I want to get good results.

I'm reminded of something a wise man once said:

"Kids, you tried your best and you failed miserably. The lesson is, never try."

 
At 2:41 PM, October 25, 2005, Blogger Kat Coble said...

Here I was all excited after the buzz around blogelo yesterday.

As you should be, really. It's a great idea that's just suffering from early-adapter manipulation. Honestly, how do you think Buzzmachine and Instapundit got their cushy thrones? They were early adapters who learned to exploit the system, while being very very good at blogging.

I've had my own Thai Food with an unnamed dinner companion and have a few ideas:

If I made it where people couldn't vote on their own sites then I'd have to figure some way to prevent them from setting up another user and using that one to promote their site.

That's viral marketing of the truest sort. I can't imagine that many people would bother going to those great a lengths. The few that would could be a problem, definitely.

One possibilty is that each succesive vote for a particular site within a certain amount of time adjusts the K down.

That could be a good way to work it, but I don' know how badly that would skew your initial variable and then render the system obsolete as a true Elo measurement.

I'm still seeing the major issue as resting primarily with the user-based voting capacity. If there is a way to limit that it would work far better. If Blogelo truly catches on in the larger sphere, I envision full-on bot abuse to skew the statistics for promotion. You'd at the very least need to have some type of word verification a la Blogger to curtail that.

I personally think this is really workable if you can iron out a few minor kinks. Don't give up. Seriously.

 

Post a Comment

<< Home