brewerfan.net header
About Us    Power 50    Link Report    Daily Brew    Draft    FAQ    Links
Transactions    Player Index    Login    Fan Forum
Lambeau Leap 1250 WSSP  
 
Weekend Report: Post-Season Odds
 
2003-08-23
Player of the Week: Scott Podsednik
 

As Brewer fans, we can only dream about the post season. But as we know, the nature of the pennant races does affect us; who is buying and selling at the deadline can have a profound impact on the franchise. It is important for buyers to know what kind of shot they have at the post season. It's even better if they can quantify how much a particular player can help them improve their odds of reaching October. Into this environment I introduce the Post Season Odds Report (link http://www.baseballprospectus.com/current/ps_odds.htm). The report lists the odds each team has of winning its division and taking the wild card, as well as a total post season percentage. The report is updated daily and at this stage of the season a lot of teams have zero chance.

The Odds Report builds on the concepts of the 3rd order winning percentage statistic I introduced a couple of weeks ago. The chances of a team making the post season are based on 3 factors: the quality of the team itself, the difficulty of its remaining schedule, and the performance of competing teams. We know how to find out the intrinsic quality of the team - it's merely the 3rd order winning percentage. This provides the base for the next couple of steps. The strength of schedule is influenced both by the teams played and the number of road or home games. Taking these factors into account makes it fairly easy to count up an aggregate winning percentage for the opposition. This process is repeated for all of the other significant competitors the team has.

With these numbers, it is relatively easy to generate a distribution of the number of games each team can be expected to win the remainder of the season. This takes the form of a simple curve with win totals as the X axis and a percentage for the peak height. For those in need of a visual please follow this link http://premium.baseballprospectus.com/article.php?articleid=2205 which gives the BP explanation of the statistic and a number of charts. In simple terms you can add the team's current record to the projection and ask what chance the team has of winning that many games. For example, at the time they wrote the article the Cardinals had a 12% chance of winning 86 games. This doesn't really tell you what your chances are, since the Cubs and Astros might win 85 or 87 games. The simple elegance of the system is that you can add up all of the various possibilities rather easily (with a computer). Start with one team at the lowest win total that's likely. In this case the Cardinals have a 1% chance of winning 77 games. You can then ask what the chances are for the other two teams both having an inferior record (that would involve adding any chance for the team to have less than 77 wins together and multiplying it by the percentage for the other team having a total that low), and in this case the total is zero. You repeat for each win total and add all of the percentages together. As of yesterday, that calculation indicates that the Cardinals have a 21.1% chance of winning the division. To determine odds for the wild card is the same, except it involves more teams and you have to remove division winners. The Cardinals currently have a 0.4% chance of winning the wild card.

I'll give a brief explanation of the various statistics BP posts in the Odds Report. They list the 3rd order winning percentage and strength of schedule (SOS). They also have the somewhat confusingly title ROY%, which is the winning percentage calculated for the team for the remainder of the year based on the above factors. It is not the winning percentage the team is expected to finish the year with! Next they have a W, which is the mathematically expected number of wins for the team at the end of the year; normal people just call it the average. Next they list the percent chances for division titles and the wild card followed by a post season total. Finally they list the Key W. That is how many wins the team needs to get to have a better than 50% chance of getting to the post season.

What can you really do with this information? It is important to realize the limitations of this method. First, it requires reliable information about the team's quality. That means that it's not really informative until June at the earliest when you can have a fair estimate on how good teams are. To do these kinds of calculations in the offseason, you need a system more like Diamond Mind baseball which simulates a baseball season based on realistic player evaluations and expectations. Second, the statistic is sensitive to violations of two key assumptions - namely the independence of the games and that the measured team quality is a good estimation of the team's true quality. First, what the independence of games mean is that a win by one team does not change the probability of another team in the race winning. We know this isn't true when two contenders face each other down. It's not a real problem unless the teams play a lot of games against each other in the latter part of the season. We'll call the second situation the DL exception. Take a team (we'll call them Arizona) that has a couple of big impact players sitting on the DL. Those two players are scheduled to return to the club near the trading deadline and haven't really contributed to the team's quality in the first half. In this case, it seems fair to acknowledge that in the second half the team will likely be better. This isn't a major problem if you can estimate how much of an improvement the players will be. In fact, that is the most useful aspect of the statistic - weighting the benefits of various player additions.

There are a variety of ways to translate a player's production into wins, but in general a superstar replacing a scrub will be about six more wins for a team for an entire year. You can go from there and figure an above-average player plugging a hole on your team at the trading deadline will net you roughly one more win in the remainder of the season. It's a simple matter to add that win into the above totals and recalculate the percentages involved. Similarly, this could be done for Arizona to adjust for those two stars coming back.

As I said earlier, they update the percentages daily because they change (sometimes dramatically in close races) as each game is played and the outcome goes from estimated to certain. I think this is a fantastic tool for judging how valuable deadline deals are. For teams in tight races for a post season chance, a key upgrade can change their chance of making it by as much as 20%! That's a lot of value. I'm pretty sure the Brewers haven't come anywhere close to having a total 20% chance of making the post season in the last 11 years.

Player of the Week

An early statement game by Scott Podsednik (4RBIs and run), followed up by a two-hit performance on Sunday definitely had him pacing the team for the week. He continued to produce at least an RBI or a run for the rest of the week, with the exception of yesterday. Others had big games during the Philly sweep, but Scott gets the nod for being there all week.

 
 
OTHER ARTICLES BY
Jason Belter:
 
Weekend Report: The Ten Run Rule
(2003-09-05)
 
Weekend Report: Pitcher Abuse Points
(2003-08-29)
 
Weekend Report: Post-Season Odds
(2003-08-23)
 
Weekend Report: Statistics - Defining Failure
(2003-08-18)
 
Weekend Report: The Return of a Classic
(2003-08-08)
 
Mid-season report: Firesale vs. Holding Steady
(2003-07-04)
 
Weekend Report: Dearly Departed
(2002-09-20)
 
Weekend Report: Cheese Platter
(2002-09-13)
 
Weekend Report: New Direction
(2002-09-05)
 
Weekend Report: Beyond the WHIP
(2002-08-30)
 
Weekend Report: Stats Demystified - ARP
(2002-08-23)
 
Weekend Report: Blockbuster Time!
(2002-08-16)
 
Weekend Report: The Brave New World
(2002-08-08)
 
Weekend Report: The 31st Corporation
(2002-08-02)
 
Weekend Report: Position Rankings for the Divison
(2002-07-27)
 
Weekend Report: Lies, Darn Lies, and Statistics
(2002-07-18)
 
Weekend Report: How do our staters compare?
(2002-07-11)
 
Reviewing some predictions
(2002-07-05)
 
Weekend Report: Bad Pitching or Bad Defense?
(2002-06-28)
 
Weekend Report: Stats Demystified - EQR, RARP
(2002-06-21)
 
Weekend Report: Stats Demystified - EQA
(2002-06-13)
 
Defense and the Brewers
(2002-01-26)
 
Rating the catchers
(2002-01-08)
 
 
 
 
Questions? Comments? Contact Brian Kapellusch (president, systems engineer) @ webmaster@brewerfan.net
Brewerfan.net is a fan-based independent site, and is NOT affiliated with the Milwaukee Brewers Baseball Club. Please support the Milwaukee Brewers by visiting their site at http://brewers.mlb.com
 
  Member Site of the Fansite Alliance