HOW DATA CAN WORK (BETTER): Dan Altman, economist and owner of North Yard Analytics LLC, has a three-part series at Bloomberg Sports' StatsInsights about statistics, analytics and soccer. Part one focused on correlation and the pitfalls of strictly linear statistical tests. Part two looked at creating probabilities within aging curves and talent level. Part three shared an example of recruiting for why analytics and traditional scouting can be very beneficial when combined together.

PLAY BALL: ESPN Insider's Keith Law looked at how team shifts are affecting opponent stats via BABIP and well-hit ball allowed stats. The Hardball Times' Max Weinstein has a really detailed look at simulating and identifying platoon players. FanGraphs' Dave Cameron wrote at FOX Sports about umpire changes that could bring down the ever-increasing strikeout rates. And Baseball Prospectus' Russell Carleton wrote about how sabermetrics still needs translation research to make a larger dent on the field.

ON NBA EXPECTATIONS: FiveThirtyEight's Nate Silver got me thinking this week: How much do expectations matter for coaches? His study this week was on how Vegas win total odds affect coach job security. The feeling is familiar to me as a Cleveland fan, where the Cavs improved by nine wins up to 33 in 2013-14, but it felt like a disappointment compared to their 40-win odds. It would be neat to see how this compares in other sports as well.

BASKETBALL PLUS-MINUS ANALYSIS: APM, RAPM and RPM have been all the rage of late. Some sites besides ESPN also are doing great analysis -- and were doing it before the recent unveil. GotBuckets' Randall Cooper shares an introduction to Four Factors analysis with APM data that will help to break down the data more directly. Hickory-High's Jacob Frankel explained how to calculate RAPM. And Austin Clemens introduced a really cool visualization into adjusted defensive impact by court location. This uses a similar methodology as APM-esque statistics.

OTHER BASKETBALL LINKS: ESPN Insider's Tom Haberstroh debunked the myth of pre-playoff momentum. Seth Partnow shared how certain traditional offensive systems, such as Houston's post use of Dwight Howard, might be out-dated and inefficient. FiveThirtyEight's Benjamin Morris wrote about the possible value of technical fouls. RealGM's Jonathan Tjarks wrote about why the eye test still matters in relation to some top NBA first-years. And Justin LeBeau wrote about about the possible "golden era" of today’s NBA.

ON THE GRIDIRON: Pro Football Focus’ Mike Clay shared some neat data on depth and passer-adjusted catch rates. RotoViz’s Davis Mattek had data on height, weight and touchdowns. If you didn’t see it yet, the Sun Sentinel’s Omar Kelly wrote about the Dolphins’ new analytics department. And impressively, ESPN.com’s NFL writers predicted 2014 records and somehow managed to have the league finish 68 games over .500 as a whole with only two teams below seven wins. Now that’s a touchdown.

NEW DATA WEBSITES: Over the last two months, we’ve seen a slew of new websites popping up promising new philosophies of modern data journalism. There’s Nate Silver’s FiveThirtyEight, now under the ESPN umbrella and doing more than just sports and elections. There’s Ezra Klein’s Vox, now under the Vox Media umbrella that also owns  SB Nation, and this site focuses mostly in “wonky” policy explainers and interesting new card stacks. Now, there are two more unveils to add to the fold at the former homes of Silver and Klein..

On the more serious side of things, David Leonhardt’s The Upshot is the new data project at the The New York Times. It will mostly also feature cool numbers and visualizations on politics and the economy, but will occasionally dabble in sports (like today). And The Washington Post also now has Neil Greenberg’s Fancy Stats, a quick sports news and numbers blog from the hockey data whiz. You definitely should keep all four of these sites on your radar screen. They have different philosophies, styles and approaches to sports, but everyone will benefit from their approach to data.