Sunday, March 26, 2006

March Madness :: Final 4 Probabilities!

I was looking on to see if anyone picked George Mason in the Final Four... and lo and behold, there was a guy who not only has them in the Final Four, but has correctly picked GMU, Florida, LSU, and UCLA!

It's pretty amazing. It seems freakish at first glance. It's one thing to pick an 11 upset in the first round, or the 2nd round. To have that upset all the way through to the final four, along with nailing the other 3 teams correctly (who are 2,3, and 4 seeds)... defies the odds. This should be worth more than $10k in my humble opinion. ESPN and Pontiac are getting a deal on this marketing!

There are very close to 2,681,679 brackets that were filled out on (each person is allowed to fill out up to 5 brackets per signup, so it's not necessarily 2.6 million people). The reason I know this is because according to their leaderboard, I am in the 94.7 percentile with a single Final Four team selected (UCLA), and my rank is a distant 142129. ...=142129 / 5.3% = 2,681,679. (surely this 94.7% will plummet in the next couple of days as I have ZERO possible points remaining with UConn being knocked out!)

The number of possible outcomes to the tournament is 2^64... which ends up being: 18,446,744,073,709,551,616 to 1 that you get all 63 games picked correctly. But that is to get every single game correct, those aren't the odds to get the final 4 teams. To just get the Final Four teams, (I'm not 100% sure my math that follows is accurate, but I believe it is: please correct me and provide justification if I am wrong) you have to get 16 coin flips correct in succession (4 teams through 4 rounds -- which would be 50%^16 = 0.001526%. But that assumes you know the four teams from the beginning, it doesn't include the selection of the 4 teams, which would be 4/64 * (50%^16) = 0.000095367%. This makes 1 in 1,048,576. So it makes sense that they have at least 1 person with these final 4 teams. logic might be incorrect though, but that's how I see it, with the teams as ping pong balls. I'd like to get the answer from a probability guru.

Anyway, March Madness... great times!

Here is the guy who correctly picked these 4 teams:
Here is my defeated bracket :(

Monday, March 20, 2006

AJAX Position Available in Redmond, WA

There is an exciting AJAX position available in Redmond, WA for interested developers. The customer is re-designing the online B2B and B2C webpage and need some front end java developers with AJAX experience. This is a 6 month contract to hire and is with a company who had over $300 million dollars in profits in Redmond, WA.

I was interested myself, but because of the project I'm on I am unavailable.

They are moving fast and the faster you can get Cheryl your resume the better. If you are interested, contact:

Cheryl L. Burdett
Technology Recruiter
Phone: 206-720-5312
Cell: 206-437-1473
Fax: 206-322-2718

Friday, March 03, 2006

Google : Admiring their top priority

As of yet, I do not own Google stock, but I have admired their company for over 4 years since a colleague at my last company introduced me to their search engine. I admire their innovation in pushing web technologies forward and their philosophy on how they develop and deliver service to the consumer. There's been a lot of misplaced media 'fluff' about "click-fraud", so I wanted to give my opinion.

Here's why I don't think "click fraud" will threaten Google at all. It's a great free market system as it self-regulates. For as long as people show a return on investment they're comfortable with, they will keep buying ads. What many not in the business (and apparently many writers) do not know, Google provides the ability to imbed pre-formatted javascript into your 'buy confirmation' webpage, that enables you to track the exact percentage of users that clicked on your Google ad and then went on to buy. It is the best form of advertising on the planet that is: targeted and measureable.

Even if there is click fraud vs a particular company, one of the effects of this is causing that particular advertisers "click through rate" to go to astronomically high levels. "Click through rate" is the number of clicks divided by the number of page impressions, and is used as one factor in determining ad position. Ad position and frequency are not determined solely by cost -- it's relevance to the keywords, click-through rate, and cost. So, this person whose ad was being assaulted by clicks would have to pay less for the top spot than another whose ad wasn't clicked on as much. If done at extreme (E.g. automated) levels, it could perhaps cause those advertisers to drop out for bidding for those keywords at that price... however these drop outs cause less demand for those keywords, lowering the bidding price and letting others in (or the original company back in). In other words, the fradulent clicks are built into the price-per-click already. It has been and continues to be a brilliant system for the consumer, the advertiser, and Google. What's great is that their first priority, even in delivering advertising, is delivering links and information that are relevant to the consumer's search request.