Posts filed under “Data Analysis”

A Modest Proposal


Now for something completely different.

That we have become a numbers (and statistics) obsessed society is beyond debate. I’m not sure when it started (Cosmo mag, maybe?), but it’s gotten out of hand. Websites everywhere tell me the X number of things I need to know about Y. Sports, particularly baseball, have gone totally over the top: “Well, Bob, Smith is pretty good in this situation: Over the course of his ten year career, when batting at home, in the bottom of the third, with runners on first and third, one out, with a 1-2 count, on a Wednesday evening in the month of July, between 8:15 PM and 8:30 PM, with the temperature between 70 and 72, he’s…1-for-1.” Much of this, of course, owes to technology and advances in computing power over the years. We seem to compile some statistics for no other reason than we can. As a numbers guy myself, I guess I don’t mind too much and would like to add my own idea to the mix.


I’ve always been a big fan of Scrabble, and a daily game or two against the computer on the iPad is now part of my daily routine. Scrabble’s a game, to the best of my knowledge, to which a meaningful new stat hasn’t been added in approximately forever. So, herewith my modest proposal:

Points per Points Played (PPP), the formula for which would be, as indicated, Points Scored (what did you score?)/Points Played (sum total of tile points you played to score it). I consider this a good measure of a player’s efficiency with his/her tiles. We know that the sum of all tiles is 187 face value. Those points are played with varying degrees of success from game to game to produce whatever outcome. A simple 1-point letter “S,” placed on a Triple Word Score box, could easily generate 30 or more points – a huge ROI for a 1-point tile. That same “S,” appended to the letter “A,” would produce only 2 points, a vastly inferior outcome. Someone who, during the course of a game, played 95 points of tiles and scored 400 points would have a PPP of 4.21. His opponent, who played the remaining 92 points and scored 410 points, was slightly more efficient (in winning while playing fewer points), scoring 4.46. PPP could also be a career statistic, allowing players to capture a reasonable measure of proficiency over time, as I’d expect better players to have a higher PPP. This measure, which could of course be calculated manually, would be a snap to incorporate into the iPad app. While there obviously other measures of proficiency, such as, say, points per turn (also valuable), PPP rewards you for efficiency and not just for playing high point value tiles such as Q, Z, J, or X.

And, yes, of course I understand that I have not accounted for the two blank tiles.

So, there you have it. Thoughts?

Category: Data Analysis, Mathematics

Have We Passed “Ex-Bubble Employment” (adjusted) Peak of 2007?

Source: FRED     Here is an interesting question: How does the present employment rate compare with the prior peak? According to Bureau of Labor Statistics data, if we compare the prior recent peak of 139,143,000 of November 2007 with last month’s figure of 137,942,000 (I used non-seasonally-adjusted because it’s November to November), we are…Read More

Category: Data Analysis, Employment

Comparing Private Job Creation Now & Then . . .

@TBPInvictus here: David Rosenberg made a point in his note Monday that I don’t think went quite far enough, or at least needs a bit more color:   Rosie then showed the chart below (the Haver Analytics version of it, anyway), which he dubbed “Employment Less Financials in Private Sector.” In St. Louis Fred-speak, that…Read More

Category: Data Analysis, Employment

Terry Jeffrey, Hack Extraordinaire (Fun with Numbers Again)

@TBPInvictus here. In yet another stunning display of journalistic malpractice, Terry Jeffrey put out the following piece after last Friday’s NFP release:     Because it suited his (political) purpose, Jeffrey jumped on the Household Survey; looking at the Establishment Survey would not have provided such a dramatic headline. Indeed, it would have provided, for…Read More

Category: Bad Math, Data Analysis, Economy, Employment, Financial Press, Really, really bad calls

Media Hall of Shame: Black Friday Reporting

I am writing up my BBRG coverage of the innumerate business reporting of Black Friday and the holiday shopping weekend. The press as per usual got it wrong again this year. Here is the press release from the  NRF, pushing their usual survey silliness as if it were actual retail sales data: “More than 141…Read More

Category: Bad Math, Consumer Spending, Data Analysis, Financial Press, Really, really bad calls

What If All 50 States Had = Populations?

Source: Fake is the New Real Hat tip GovBeat

Category: Data Analysis, Digital Media

Category: Data Analysis, Think Tank

This is how we die now…

From the Washington Post: Every year, premature death — that is, death attributable to causes other than old age — deprives people of a combined 1.7 billion years of life they could have enjoyed. 1.7 billion years. Think of everything those people could have done — the children they could have had, the jobs they…Read More

Category: Data Analysis, Digital Media, Markets

Has the Dollar Really Lost 97% of Its Value? (No)

  One of the favorite tropes of the “End the Fed” crowd is the “falling purchasing power of the U.S. dollar.” Google that phrase, and you will be rewarded with 91,100,000 results. (drop the “U.S.” and it doubles to 187,000,000 results). The problem is, nearly all of these arguments are wrong. As Matt Busigin of…Read More

Category: Currency, Data Analysis, Really, really bad calls

Interactive Data: Bloomberg Industry Leaderboard

Supercool use of data from the DataViz team at Bloomberg BusinessWeek

Category: Data Analysis, Digital Media