Predicted Wins and Playoff Probability: Week 12

It has been awhile since I last posted the Orioles predicted wins and playoff probability.  All of the systems were sliding between 76 and 80 wins.  That just did not seem an incredibly exciting thing to update every week.  So where has a month later left us?  Between 76 and 80.  Perhaps it is slightly exciting to note that the Camden Depot and PECOTA predictions are beginning to move away from the Pythagorean.

A few more detailed notes:

• The Orioles took four wins against Seattle during the six games in which they will face each other this year.  My system predicted 3.43 wins, so this was a good outcome for the team.
• They split even with the Nationals three games a piece.  3.42 wins were expected.
• Against Cincinnati they took two of three while the expectation was 1.47 wins.
• They dropped two of their three to the Pirates where they were expected to win 1.67.

Out of these four finished series, they are +0.01 wins based on my initial projection.  This is practically breaking even.  Series against the Yankees, Jays, and Indians are not going so well and the team is likely to under-perform against those teams.

Before we get to the updated graph, here are some solid posts over the past month:
They Lathe Bats, Don't They?  An interview with a guy who makes bats for some of the Orioles
Science of Baseball Always a crowd pleaser
Revisiting MLB Reallignment

Click on image to make larger

Predicted Wins and Playoff Probability: Week 9

The Orioles were involved in two sweeps this past week.  They took the Royals to task during the weekday series finishing with a three wins.  However, things skid to a halt (including a rare dismantling of Zach Britton) in Oakland where they had three straight games taken from them.  In response, they have junked three-fifths of their rotation with Brad Bergesen and Chris Tillman being relegated to Norfolk where they will undoubtedly dominate AAA batters.  If they are not in the rotation, they are probably best served in the MLB bullpen.  I imagine their stay will be short.  Brian Matusz will be pitching for them on Wednesday.  It is assumed that Alfredo Simon will be given a shot at the five slot in the rotation.  I should have made sure you were sitting for that.

Anyway, the projection models are hovering around the 78 and 79 wins with the changes becoming less significant as you can see in the graph at the end of this post.  The team will need to finish 69-42 to get to 93 game won and a likely playoff berth.

Here are also a few links from the past week that might be of interest:

Nick's summary of potential Baltimore Oriole targets in the first round.
The latest thoughts of mine for wishing to do away with collisions on the base paths.
A post on some less discussed players of interest in the later rounds of this year's draft.

Predicted Wins and Playoff Probability: Week 8

The Orioles have been interchanging between good weeks and bad weeks.  This is reflected in the predicted wins displayed in the chart below.  The magnitude of the differences is getting smaller and smaller with each progressive week, but that could certainly change if several of the players get hot.  However, it remains unlikely this team is a playoff caliber squad.  In the remaining 117 games, the Orioles will need to win at a .615 clip to reach 93 total wins.  That should be good enough for the wild card or an outside chance for a low win total of a AL East champion.  The Depot has the probability at 0.17%.  PECOTA is much more optimistic at 2.6%.

Be sure to check us out on Twitter: @CamdenDepot or Facebook (be sure to friend us as we occaisionally hand out tickets we are unable to use).

Notable Post of the Last Week:
The Science of Baseball: May 22, 2011
Updated Composite Draft Board
Danny Hultzen Scouting Report

Predicted Wins and Playoff Probability: Week 7

The Orioles are quite the streaky team this year.  The last week has seen the Orioles sweep the Mariners in a three game set and going two and one against the Rays.  All three projection systems are still in line with each other, which is quite impressive to me.  Eventually, I expect a shift to occur.  I do not use the same system as PECOTA or runs given/earned.

Remember when thinking about playoffs, we are talking about somewhere around 92 wins.  The Orioles are currently far off that pace.  PECOTA puts the Orioles as having a 5% chance of hitting that.  My system has them at 1.3%.

Predicted Wins and Playoff Probability: Week 6

This past week, the Orioles were 1-6, managing a win against the Kansas City Royals.  Each projection system thought the Orioles would do better than they did and so we find the new predicted wins to be less than they were last week.  However, it is about in line with where the predictions were two weeks ago.  PECOTA has the Orioles at a 1.8% chance of making the playoffs.  We have them at 0.41%.

click on image to make larger

So, once again, this team is about as good as the systems thought they would be.

EDIT: It appears that we are getting considerable traffic from ESPN on this post.  I think I might need to explain some thing here as this is one of our weekly updates.

We are using three "systems" to predict the number of wins the team will win over the course of the season.  The three systems we are using are:

PECOTA - Baseball Prospectus issues updated projections on a daily basis.  We use this as an industry standard.  These figures are based on updated PECOTA player projections, playing time projections, and strength of schedule.

Camden Depot - At Camden Depot we used Tango's MARCEL projections and account for playing time to generate a WAR prediction.  We then generate a prediction for each team in baseball and weight the Orioles schedule according.  Expected wins are derived using Bill James log5 method.  The MARCEL projections are not updated, but the Orioles record in past games is incorporated in what we expect them to do in future games based on the projections and strength of schedule.  We will change our projections if major trades or injuries occur.

Pythagorean - This method is the simple well known Pythagorean method where we predict wins based on runs scored and runs given.

Playoff predictions are done by PECOTA in their daily update and by us.  We use a simple binomial function and adjust it to the expected Wild Card win total or AL East win total (which ever is lower).

Feel free to ask more questions.

Predicted Wins and Playoff Probability: Week 5

At the end of 5 weeks, the Orioles are not wildly over or underperforming based on the initial predictions of Camden Depot and the PECOTA forecasting. The Depot has seen a bump of 0.8 games and PECOTA is showing a drop of 0.4 games. Roughly eyeing it, I would say 5 games would be a major shift. We are not seeing that here. There have also been no surprises so far based on what my projections indicate for the rest of the AL East. Although, the Yankees and BoSox have flipped to first and second, respectively.

click image to make larger

PECOTA has the Orioles with a 5.5% chance to reach the playoffs. I have them at 1.89%.

If you prefer odds, that would be about 1 in 18 for PECOTA and 1 in 52 for the Depot.

Predicted Wins and Playoff Probability: Week 4

After four weeks into the season, consisting of 20 games, the Orioles find themselves roughly where I imagined them to be: winning about half of their games.  I did not expect them to be so streaky, but it is not surprising.

All three predictions are hovering around each other.  They are who we thought they are.

PECOTA playoff projections have decreased to 2.1%.  They were at 11.6% after two weeks.  Depot projections are at 0.71%.  After two weeks, the probability was at 3.9%.

Predicted Wins and Playoff Probability: Week 3

Quick post tonight . . . before tonight's game, the team sat at 4.2% playoff chance with PECOTA's predictions and 1.0% with the Depot's predictions.

Here is the projected wins by date:

I find it somewhat remarkable how similar the different lines are.  So far, the Depot model has had the least elasticity while the Pythagorean, as expected, has the most.  Hopefully, next week finds the team in better shape.

Predicted Wins and Playoff Probability: Week 2

There were some problems in last year's projections and probability.  I have sorted out the errors in some of the equations.  The original probability was based on the unadjusted Depot win prediction, but that has been corrected to the adjusted wins.

Click to make larger.

Updated AL East Predicted Standings: 04/09/11

In the rush of posts in the past two weeks, I neglected to put out the initial week's projected standings.  These are calculated by estimating runs scored and given by each team in a neutralized setting.  This is done by using the MARCEL projections along with a simple tool that predicts Wins Above Replacement by inputting playing time, on base percentage, and slugging percentage.  Once this is calculated I apply Bill James' log5 win expectancy model and the team's remaining schedule.  As with the expected wins piece on Monday, this will eventually take the shape of a graph when I get more data points.

03/31/11 (prior to opening day)
Red Sox 94.7
Yankees 90.0
Rays 82.3
Orioles 79.5
Blue Jays 73.9

04/09/11
Red Sox 93.3
Yankees 90.7
Orioles 81.7 (up from 4th)
Rays 80.1 (down from 3rd)
Blue Jays 74.8

As you can see a week full of games has very little impact on these predictions.  The greatest difference is the Rays who lost 2.1 wins due in part by their losing and also due to Manny Ramirez retiring.  He was on the ledger as contributing 2.5 wins this year.  However, his replacements (Casey Kotchman and eventually Desmond Jennings) account for a 1.5 decrease over the rest of the season.

2011 Season Wins and Playoff Odds: Week 1

There are only two data points, so there is not much to say.  However, the three wins nearly doubled our probability for the Orioles to make the postseason.  This is a result of the Orioles sweeping the Rays whereas the model thought they would win one game, maybe two.  The Pythagorean method is being swamped by only have three data points and no foundation to base its earlier 81 default.  It should settle down more so in the next few weeks.



Click on image to make larger.



2011 Season Wins and Playoff Odds: Week 0

Bald Eagle Hatching from Ed Smith Stadium

So begins another year and another season of keeping track of projected wins and playoff odds for the Baltimore Orioles.  In years past we have seen the three measures I use (PECOTA, Pythagorean, and Camden Depot's house stats) follow the team and mosey toward actual reality.  Typically, the team does about six to ten games worse than the projections each year.  Hopefully this season is different.

Graphs depicting the journey will appear from now on every Monday.  It being the first week, we only have one data point.  As a brief reminder, here is a rundown of each system:
PECOTA - I use Baseball Prospectus' projections that models the season and predicts wins and playoff opportunity.
Camden Depot - I used MARCEL projections to determine team WAR, adjust the WAR with a log5 strength of schedule factor, and then use a simple binomial function that determines the likelihood of the Orioles winning 92 games.  Why 92?  Based on all of the projection systems I have seen, 92 wins is likely what will result in a wild card.
Pythagorean - It assumes that what has happened will happen again no matter the sample size.  I will also use the same binomial function as a playoff predictor.

Projected Wins:
PECOTA - 80.3 wins; 8.20% playoff chance
Camden Depot - 79.5 (80.7 without adjustment); 2.23%
Pythagorean - 81 (by default); 3.56%

The Depot's current projection of 2.23% is currently the lowest of the systems I can find.  CAIRO puts them at 12.5%.