Last week, my odds picked a final four of Loyola, Notre Dame, Maryland and Colgate. Three out of four ain't bad. At the end of that article I picked against my odds in a couple games and went with Loyola, Virginia, Duke and Maryland. Again, three out of four. This week, I'm hoping to pick the national title game correctly.

As always, these odds are based on the Log 5 method, which is equivalent to the Odds Ratio. I have two models, both are possession-based and work off of offensive and defensive efficiencies. One model is raw data that is adjusted for schedule. The second model uses regressed shooting rates. Those are then used to create regressed offensive and defensive efficiencies and from there I adjust for schedule faced. A simple average the two Pythagorean Win Expectation's give me my overall rankings which I use in the odds that I'll present below.

With that said, your odds for Championship Weekend are listed in a nifty little bar graph.

You can click the image to enlarge.

Continue reading this post »

In my odds last week, I wrote that Massachusetts was the favorite to win the entire tournament and that the three upsets I'd take were Maryland, Denver and Princeton. So, I was right on two of my game predictions and the favorite has already been ousted from the tournament.

I like it.

In the first round, teams I had favored by my rating system won all but three games. Massachusetts had a 61.3 percent chance of moving on, Virginia was given a 34.5 percent chance to beat Princeton and Lehigh was given a 51.4 percent chance to beat Maryland. Not bad. Then again, just going by seeding, there were only three upsets as well. Enough of that. Odds for the remaining teams have been calculated.

Quick note, as always: this data is derived from my own tempo-free statistics which combine two different models. One is based on raw offensive and defensive efficiencies and adjusted for schedule. The second model is the same but uses regressed shooting rates to form regressed offensive and defensive efficiencies. Then, I adjust for schedule using those regressed efficiencies.

Math malarkey out of the way, your odds for the eight remaining teams in the National Lacrosse Tournament are encapsulated in the following graph.

You can click the image to enlarge.

Continue reading this post »

Unfortunately I ran out of time to continue the conference tournament odds last week, but I followed the same template and came up with odds for the national tournament.

As always, these are built off of my own efficiency rankings -- much like HoyaSuxa's or the data found at the invaluable Tempo Free Lax -- and combine two different models. One model is just raw efficiency ratings that are adjusted for schedule and the other model starts with regressed shooting rates, turns those into efficiency ratings and then adjusts for schedule. I average the two Pythagorean winning percentages that my models churn out and use that final average as their overall rating. Whether this is the best route or not is definitely up for debate, but it's what I'm rolling with right now. I did not include home-field in these calculations mostly out of laziness, to be quite honest.

That aside, who is the favorite to win the National Title using the Log 5 odds method and my computer?

Continue reading this post »

Eds. Note: I missed this yesterday from Mike. Front page'd because my eyeballs are starting to cross and fingers starting to bleed.

I say Big East and you think "Syracuse, duh." This year, however, the correct answer has been Notre Dame. The Fighting Irish won the Big East conference and, as is customary in sports, are awarded the top seed in a conference tournament. Plus, they get the baller suites in stunningly gorgeous Villanova, Pennsylvania. (I've never been there, nor have I known anyone to ever go there, so I can only assume it's stunningly gorgeous.)

As always, these Log 5 ratings are based on my own efficiency ratings like HoyaSuxa used in beating you over the head with Big East tempo-free team profiles of each of the participants. Also, RyanMcD29 has some great, pretty -- far prettier than Villanova, Pennsylvania, I'm assuming -- infographics on the tournament for your viewing pleasure right here. My efficiency models are a combination of regressed and un-regressed shooting rates. Onward and upward.

You can click the image to enlarge.

Continue reading this post »

I'm a bit late with the CAA -- err, I mean THUNDERDOME! -- conference tournament odds, but it's better late than never, right? If you want to read about each team more in-depth, all of your THUNDERDOME! goodness is encapsulated right here in this very link. Really, what more could you want? Pretty infographics, enough data to make your eyes bleed, some humor. You know, besides some Log 5 conference tournament odds.

As always, these are presented on the backs of my own efficiency-based ratings like the ones in the above link or the ones found at the great Tempo Free Lax. Mine combine a regressed and un-regressed model to arrive at a final Pythagorean win percentage. And away we go.

You can click the image to enlarge.

Continue reading this post »

Eds. Note: In the haze of tens of thousands of words on the ECAC yesterday, this flew under the radar. To the front page!

The Patriot League tournament worked out much better than the ACC Tournament did with regards to my odds posts (though, I still had Colgate as the favorites to win the game against Lehigh, it was pretty close to a coin flip). Now, the ECAC Tournament kicks off -- as you can tell by the mounds of information for your viewing pleasure College Crosse boys have whipped up for you here -- I bring my goodies to the information feast. That is, of course, tournament title odds!

As always, numbers are based on a mixture of my two tempo-free models. One is using regressed shooting rates and one does not. Both are adjusted for schedules faced, then I get an average Pythagorean winning percentage which I use in these Log 5 tournament odds.

Let's jump right into it. Your odds graph looks as such:

You can click the image to enlarge.

Continue reading this post »

Ed. Note: To the front page! Remember, folks: Instead of leveraging yourself deep into soy futures, think about an investment strategy that includes taking Bucknell and the over.

Much like the odds I ran for the ACC Conference Tournament (those turned out fantastic, didn't they?), I come bearing odds of each team winning the Patriot League conference tournament title.

My ratings are based on the same theories as the team-by-team profiles Sir Hoya Suxa presented earlier (all of your Patriot League knowledge can be found here; use it and impress your friends tomorrow when everyone you know gathers 'round to watch the tournament) and the numbers found at the awesome-sauce Tempo Free Lax.

The one difference is that I have two models, one using raw data adjusted for schedule and then one that uses regressed shooting rates and then adjusts for schedule. I average the two ratings to get an overall snapshot at each teams talents.

Without any more stalling, your Log 5-based Patriot League Conference Tournament odds look as such:

You can click the image to enlarge.

Continue reading this post »

Ed. Note: Front paged, but proceed with caution and certainly don't use such information for gambling purposes. (Unless you're putting your kidney on the line. Then, yeah, I whole-heartedly recommend use this information for gambling with your life purposes.)

With Duke's beat down of Virginia last week, the Blue Devils locked up the number one seed for the Atlantic Coast Conference tournament. That was good enough to draw them fourth-seeded Maryland. The Cavaliers will get the North Carolina Tar Heels in the first round as well.

Duke, by seeding, is the favorite. But is that actually true? First we need baseline talent levels for each team.

I have two models that are efficiency-based like the results that you see at Tempo Free Lax or when HoyaSuxa presents his data. One of my models is raw efficiency data that I adjust for schedules faced and the other uses regressed shooting for each team and then adjusts for schedules. I've averaged the results of these models together to give me my Pythagorean win percentages. My odds for each team getting to the championship game and winning it are. . .

You can click the image to enlarge it.

Continue reading this post »

Ed. Note: Scrubbing out Wagner's first win since April 2010? Yup, that's front page-worthy. Enjoy!

When a team goes winless over the span of 721 days, it's noteworthy. Thus, when a team breaks their win-less skid that lasted 721 days, it's even more noteworthy. Sir HoyaSuxa did a fine job writing up a post about Wagner's first victory of the year -- don't worry Sacred Heart, someone was bound to drop a game to the Seahawks -- but I wanted to take a deeper look in a quarter-by-quarter breakdown of the landmark victory for a mostly hapless lacrosse team.

Going into the game, Wagner had about a 12% chance of winning against Sacred Heart by my calculations using my efficiency ratings. 10% of that came from being on home field. Needless to say, they were out matched. One of the ways I like to look at individual games is using my version of a Four Factors graph. In basketball -- especially college hoops -- there are what are called the Four Factors. These four statistics are what contribute most to winning basketball games. Intuitively it makes sense. In looking through the things that contribute to winning lacrosse games at the team level, I've come across four main components*:

1. Offensive/Defensive Efficiency
2. Possession Percentage
3. Turnover Percentage
4. Clearing Percentage

Here's what the factor graph looked like for the Wagner game as a whole:

Click (all) image(s) to enlarge.

Continue reading this post »

Ed. Note: Good work here out of Mike. To the front page!

via images.lax.com

After a 13-11 defeat in Durham to arch rival Duke, Tar Heel head coach Joe Breschi decided it was time to change up the lineup. He moved stellar freshmen Jimmy Bitter and Joey Sankey into the starting lineup and paired them with junior attack Marcus Holman. The results were three straight wins over ACC foe Maryland, Dartmouth and then-top-ranked Johns Hopkins by a total of eight goals.

Then they were thrashed on the ESPN1, The Uno, in front of what I'll assume were hundreds of millions by the now-top-ranked Virginia Cavaliers. What has changed for North Carolina since moving a couple of fab frosh into the lineup?

Continue reading this post »

I wrote about this once and after I hit publish and thought more about it, I didn't like the results of my first effort to try to quantify defense with simple box score stats.

So, I tried to come up with the best way I could to take the defensive ratings from Sports-Reference's College Basketball Site and turn them into a defensive version of KJ's PORPAG. All data is from either Friday or Saturday; don't remember when I grabbed it, exactly.

Continue reading this post »

Edit: for what ever reason it never fully occurred to me that the defensive usage rates should be summing to 100%. That's an oversight and error on part that I'll fix at some point.

I love PORPAG. I've written about it here a couple times. It's a great snap shot at a players talents, efficiency and how they're used in their teams offense. What it doesn't do is incorporate defense. Well, I'm going to step into the bear trap that is presenting defensive ratings based on simple box score statistics.

Ideally, we'd have 345 stat nerds watching a team's every move on the court and charting what they see, like David Hess started doing last year (and brought to Team Rankings) and the Duke Hoop Blog does. This produced great results from a collaboration between Luke Winn and David Hess for Sports Illustrated. Unfortunately, we don't have that. What we do have are boxscore stats. We also have the wonderful Sports Reference College Basketball site who have Defensive Ratings for each player for since 2009-10. This is the result of about 90 minutes of fooling around in Excel.

Continue reading this post »

Last March I had three separate posts on regressed shooting rates for Big Ten basketball teams. Since I was toying around with some numbers I figured I'd throw it into a FanPost for all to enjoy.

Essentially, I'm figuring out how much regression is needed for shooting rates among the 12 Big Ten rates using the method Tom Tango laid out in this blog post. This is kind of an update on the first of my three posts last March.

Continue reading this post »

Ken Pomeroy release his ratings for the 2011-12 season the other day. So, I thought it'd be interesting to take a quick look, in graphic form, at how the MAC is shaping up in a tempo-free light.

Continue reading this post »

These are presented with out much commentary as I'm pretty busy and the Wings have another game in about two hours. I will say that I don't think that the Red Wings played that bad. A few defensive miscues, some good breaks for Washington, and special teams goals all adds up to a lopsided loss to a team that isn't that much better (if at all) than the Wings.

Continue reading this post »

Columbus came into Detroit as the lone win less team in the NHL and left on Friday night as the lone win less team in the NHL. The play at even strength was pretty even, but the Red Wings separated themselves on the power play. Columbus didn't generate a single power play scoring chance that I tracked tonight while Detroit had three chances (more "opportunities" which either didn't get to the net, were from too far out to really be considered a "chance" or weren't able to even get a shot away.) and seemed to capitalize on many of their chances regardless of game state.

As always, big ups go to Vic Ferrari of Time On Ice fame and his wonderful script allowing me to present these numbers.

The Red Wings numbers are in red (hooray backround's with different colors!) and Columbus' are in blue. Detroit only out-chanced Columbus by one at even strength, 7-6 but held a 13-8 advantage for the game on the back of their three man-up chances -- two of them coming on the 5-on-3 that Nick Lidstrom eventually scored on.

Continue reading this post »

I've finally had the chance to catch back up with the Red Wings games (just in time for a back-to-back set on Friday/Saturday). If you'd like to re-read the recap of Detroit's 3-2 overtime win over the Minnesota Wild -- and who wouldn't? -- that's here. The beloved CSSI analysis is right here as well.

The team totals were as such:

Continue reading this post »

Better late than never is what I always say. After a crazy weekend in which I attended the Grand Rapids Griffins home opener on Friday -- very thorough write-up of it by Jason Kasiorek here -- (and spent the two subsequent days recovering from), I've finally gotten the scoring chances tracked for last week's 2-0 win over the Vancouver Canucks. If the CSSI stats tickle your fancy, they can be found --penned by JJ, of course -- here.

Before I start, a nice stick-tap to Corey Sznajder of The Shutdown Line, a Carolina Hurricanes blog. For whatever reason, I set the DVR to record the game at the wrong time and missed the first 12-or-so minutes of the game. Corey's tracking chances for the 'Canes this year and was kind enough to track the chances for the portion of the game that I missed. Even if you're not a 'Canes fan, I'd ask you all to head on over to Corey's blog because he's a machine with very good, in-depth articles and analysis. Plus, he's an all-around good dude.

And with that, the scoring chances!

All of the Red Wings chances are in bold.

Continue reading this post »

We're continuing the Scoring Chance Project with the recap of the Detroit-Colorado game from Saturday. You can read the WIIM recap here and the CSSI analysis here. If you need a quick little run down, my recap of the scoring chances in the game against Ottawa last Friday can be found here through the wonders of the internet. I also have a wonderfully nerdy (and equally exhilarating) surprise at the end of the post. Firstly, the period totals:

Continue reading this post »

I've volunteered to join Derek Zona's Scoring Chance Project and will be keeping track of the Red Wings scoring chances throughout the season.

Firstly, the scoring chances are based off a rough, 'home plate-like' area in the offensive zone. The zone is draw like so, thanks to New Jersey Devils' SBN blog, In Lou We Trust:

Now, my guess is that early in the season the numbers won't be as clear-cut and, well, good. I don't know how many of you have attempted to track scoring chances for a game, but the amount of subjectivity is pretty astounding and I found myself drifting off into thought about what needed to be different about the play that just occurred that I did/didn't rule as a scoring chance for me to make the opposite decision of the one I made. Then I scrambled to rewind my DVR back the 45 seconds I missed.

Once I've got probably ten games or so under my belt, I'll feel more comfortable about my data, but for the first few weeks, think of these are pretty good, but probably not great, snapshots of what happened. If anything, I'm probably over-recording the number of scoring chances that occurred, but I'm not sure yet.

All of this is 100% made possible by the one, the only Vic Ferrari who has a wonderful app at his Time On Ice site that we just plug our data into and it spits out these nifty charts that I'll present below the jump. Onward!

Continue reading this post »

I think it's only fair to say that there will be some math geekery in here, but I invite all readers in, as you can look beyond any of the math and see the results after the "so what does this all mean?" header.

One of the things that interests me in sports are team shooting percentages. On an individual level, we often see a players shooting percentage fluctuate from year to year and the usual narrative is that he is distracted by potential contract situations or he hasn't been 'feeling it' recently or he didn't focus enough on it or the coach hasn't been getting Player X the puck/ball/whatever enough in good positions -- whatever becomes the easiest thing for journalists to churn out to meet their deadlines. This is true, too, at the team level.

For instance, in the NHL in 2009-2010, the Colorado Avalanche had an unexpected birth into the postseason when it was expected to be another run-of-the-mill rebuilding season. In fact, in 2008-09, Colorado finished 28th in the league with 69 total points. The next year, however, they had a season of good of netminding even though they were heavily out shot and finished 13 games over .500 and had 95 points. Those who knew what to look for -- particularly unsustainable shooting percentages for or against while not controlling the play -- saw a major step back for Colorado in 2010-2011.

How did Colorado finish this year after having a year of unsustainable save percentage two years ago? Well, they were 14 games under .500 and finished second-to-last with 68 points.

Continue reading this post »

I have always been a casual college lacrosse fan, never watching the sport until Memorial Day weekend, but it always intrigued me. This year, I found myself watching some more regular season games than I ever had and heading into the national tournament, wanted to learn more about the sport.

I've been a follower of Sabermetric stats in baseball for around five years now, and have done some writing for other SBN blogs like Beyond the Boxscore and the Tigers-centric SBN blog Bless You Boys giving my advanced-stat slant on things (and have done some tempo-free basketball fanposts over at Michigan State's SBN blog, The Only Colors). So, I'm a complete numbers geek who tends to learn more about a sport through the data mining.

In wanting to learn more about college lacrosse I went on the hunt for advanced statistics -- I knew enough that a per-possession-based stat would tell me more than a per-game rate for obvious reasons that HoyaSuxa has explained numerous times on here. As of a couple weeks ago, I started estimating possessions and tempo-free data before stumbling upon HoyaSuxa's work and now I'm hooked.

All of that was a long way to just say a couple things: 1) I am relatively new to the sport, but have a semi-decent understanding of strategy and most rules and 2) I love to toy around with statistics.

One thing that baseball doesn't lack is the ways in which you can present data. I like to apply a particular method from baseball into basketball and now I'm going to do it with lacrosse.

Continue reading this post »

A few weeks ago, Chris Jaffe wrote a nice piece over at the Hardball Times about the best games since 1950 by Hall of Fame players. (It was actually a two-part piece as part two focused on the worst performances in a single regular season game by an HOFer.) I enjoyed the articles greatly and thought, "hey, wouldn't that make for a fun article idea I could totally steal and think about taking sole credit for!" Fortunately, Chris and the Hardball Times website are both so great, that I decided to give credit where credit is due.

The stat that Jaffe decided to use was the ever-glorious Win Probability Added (WPA). In the Sabermetric circles, a large, large chunk of what is done is trying to neutralize or take the context out of the game to isolate what a player's true talent may be. This is valuable work (and a bevy of it I find highly enjoyable) as it helps to give guesses as to what a player could do in the future. For instance, using a statistic like Weighted On-Base Average (wOBA) is to properly weight each outcome of a plate appearance for a hitter to, it treats ninth-inning, game-winning long balls the same as a first-pitch homer to kick star a ball game. On a season-long level, this is the proper thing to do. However, we know that the late-inning home run is more valuable given the timing. That is where WPA steps in.

What WPA does is credits or debits a player for their positive or negative contribution in a game depending on the situation. Theoretically, you can move the needle so-to-speak anywhere from .02 to .95% in a game. Those are rare, however, and most plays wind up in the middle. WPA is calculated using Win Expectancy which, upon studying thousands of games results in baseball history, we can see, on average, how often a team won the game from a particular situation. For instance, you can consult the Win Expectancy chart and see that if a home team is up to bat in the bottom of the 7th inning and trailing by a run, the home team has rallied to win the game 35.3% of the time.  Win Probability Added uses these values to attribute credit or blame to a pitcher or hitter.

Let's run through an example using numbers just for illustrations purposes. Miguel Cabrera is leading off the bottom of the 7th and the Tigers are losing by a run. Currently, the Tigers possess a 35.3% chance of winning. Let's say Cabrera clubs a homer to tie the game. If you scroll down the WE chart linked above until you find a box for the 7th inning with the score set as "0" and no one out, the win expectancy for the home team -- the Tigers in this case -- is at .58.6% Cabrera's home run added 23.3% to the Tigers WE. Cabrera's WPA for the game would currently sit at +.233 (WPA is always expressed in decimal form). You get the difference between the team's WE at the start and the end of the players plate appearance and that belongs to him.

What is fantastic about WPA is that it's such a descriptive statistic. Baseball-Reference logs WPA and so does Fangraphs.

Continue reading this post »

A little after the mid-way point in April, I penned a piece looking at Jim Leyland's bullpen usage through the lens of two advanced statistics: Leverage Index (LI) and Expected Fielding Independent Pitching (xFIP). I used this data to create (what I thought was) nifty graph that gave you a quick over view of who was getting the most important innings and who was pitching the best, independent of his fielders.

Since that article, I've changed my methods slightly. Instead of having the number of batters faced on the vertical axis, I've now made that the average leverage index for when the pitcher enters the game -- referenced as gmLI on Fangraphs and in this article. The horizontal axis is continuing to use xFIP but now the size of each bubble is the number of batters each pitcher has faced. Let me know if you like the change or prefer my last graph for these (hopefully) monthly bullpen updates. The graph and commentary is after the jump.

Continue reading this post »

On Friday night, Justin Verlander notched his 1,000th strikeout in a Detroit Tigers uniform while pitching the Tigers to a comfortable 9-3 shellacking of the, ahem, rival White Sox (cram it, HawK!)

In doing so, Verlander became just the 15th Tigers pitcher to rack up 1000 or more K's while wearing the Old English D -- a number that, admittedly seemed low when I first heard it. Magnificent Mickey Lolich leads the list by 699 K's over second place Jack Morris. Peep the table:

Continue reading this post »

At the end of last April, I took a look at Jim Leyland's bullpen usage in a somewhat unconventional way. I used leverage index -- found at Fangraphs -- to see where and when he was inserting his top relievers. Overall, however, I wasn't a big fan of the graphI used (which was the Tigers version of a graph made by Jeff Zimmerman over at Royals Review).

An idea popped into my head and after putting it down on pape ... er, into Excel, I think I've improved on my post from last year.

What I've done is taken each reliever's Expected Fielding Independent Pitching (xFIP) and plotted that against each pitchers total batters faced this year. Then, creating a bubble graph, I used each players average leverage index when he enters the game (called gmLI on Fangraphs) to dictate the size of the bubble.

Before we present this years data, let's make a test run on the 2010 season first. Data and graph after the jump.

Continue reading this post »