I'm not sure how useful this is, but if anyone was wondering what the odds were of the "Old Rajai" batting as well as he is now without having improved at all, but by just being lucky, I did some quick math.

Assuming he is a .265 batter (his career BA) then there is a 5.6% chance he would have at least 20 hits over the 57 AB's since the break.

Assuming he is a .288 batter (his '09 BA thus far) there is an 11.7% chance.

Assuming he is a .257 batter (his first half BA) there is a 4.1% chance.

Continue reading this post »

Who would you rather have in your lineup, a fast player who can extend many a single to a double or a slow player, who regardless, has a the same SLG as the fast player?

Continue reading this post »

With all of the talk of the fast approaching trade-deadline everyone is starting to compare our players with other players in the hopes of acquiring some new talent ,and that's great. However, the way in which it is being done seams mostly based around comparing OPS, for batters at least, and quite frankly, OPS is dumb. Sure its better than batting average, and sure it takes the two most important components of hitting, how often you a batter reaches base, and how many they reach in the process, but it is very sloppy and has two major problems:

1. Slugging percentage is inappropriately weighted in terms of their actual run values for singles, doubles, triples...

2. As mentioned in Moneyball, OBP gets much less credit than it deserves as OPS is dominated by slugging percentage. (OBP should be doubled for OPS to be more accurate.)

Basically, OPS is flawed. It doesn't inherently mean anything, it's just two stats put together. It's not terrible, but why use a statistic you know is wrong? And certainly, when you're trying to predict the moves of Beane, it is foolish to use a statistic that he said was wrong back in 2004. I know its nice using a statistic that everyone already knows, but thats what they said about batting average. If its any indication at all that OPS is outdated, just look at the media, even they are using it.

So, what should be used instead? There are tons of statistics to choose from and they all have their reasons, but for everyone who wants a simple transition out of OPS, wOBA is basically a properly tuned version of OPS, and for convenience sake, it is pegged to OBP, so if you're wondering what a good wOBA is, well, its the same scale of OBP so you don't have to explain what the number means.

I am an A's fan, but this season has been impossible to watch. The offense, which was supposed to be the backbone of the team has been underpreforming epically.

OPS wise, the offensive numbers are as follows from 2008-2009 for the nine most frequently batting players this season:

<!-- @page { margin: 0.79in } P { margin-bottom: 0.08in } -->

Ryan Sweeney: 733-643

Kurt Suzuki: 716-828

Jack Cust: 851-794

Bobby Crosby: 645-664

Mark Ellis: 694-503

Orlando Cabrera: 705-584

Jason Giambi: 875-659

Matt Holliday: 947-734

Travis Buck: 723- 545

That's a net loss of 944 between last year and this year, or an average of 104.8 a player, it's like the team gave away an all-star DH. Obviously certain players could have been predicted to preform worse from last year to this year such as Jason Giambi and Matt Holliday, but could a drop of 216/213 OPS really have been predicted, or is Billy Beane loosing his edge?

Anyway, I just wanted to see if anybody thought that this is just some beginning of the season underproduction or if this catastrophe of an offense could have been avoided.

I recently checked around the internet to find mathematical evaluations of team Batting Average and On-base percentage plus slugging percentage in determining team run totals, to no avail, so I figured I would do it myself. Because of time constraints I only pooled two seasons of data, 2004 and 2008, but that is more than the minimum 30 required to do the test. I started by finding the linear regression of both BA and OPS as dependents against runs:

'04 BA vs Runs:

Regression Line= -417+4422X                                R=.869        R^2=.475

'08 BA vs Runs:

Regression Line= -1049+6873.786X                     R=.803         R^2=.646

'04 OPS  vs Runs:

Regression Line= -986+2315X                                R=.971        R^2= .943

'08 OPS vs Runs:

Regression Line= -730+19971X                              R=.948        R^2=.898

Totals: BA R=.746  BA R^2=.561; OPS R=.96  OPS R^2=.921

* Because I did this on my calculator and not on Excel I don't have the graphs available to show, sorry.

From here, we can easily see that although BA does predict runs, it is not nearly as significant as OPS which does so at an outstanding rate.

If we were to use hypothesis testing to evaluate this claim we would find that the Test statistic for BA is 8.861 compared to 26.894 for OPS, both over the approximately 2 critical point, causing us to fail to reject both, although showing how much more confident we can be in OPS as a predictor.

In conclusion, it is my analysis that although batting average does have a correlation with runs it is not a particularly strong one, especially compared to OPS which, at least in 2004 and 2008 had a very strong correlation to runs produced.