Simulate The Wildcard Race With Excel

[rluzinski]

For the 3 people who might like to try this:

 

http://www.mediafire.com/...57cc11156d2db6fb9a8902bda

 

 

Edit: Spelling correction in title.

[Oldcity]

Bah. Filtered out at work.

[rluzinski]

PM me your email and I will email you the file, if you would live.

 

To use it, first choose the skill level type you want to use:

 

WIN%: Assumes a team's win percentage perfectly reflects the true, collective talent of a team.

 

WP3: Baseballprospectus's estimate of true talent.

 

If you want to adjust the skill levels, select WP3 on the schedule sheet and edit the values in the WP3 column on the win% sheet. BP says the Brewers are a true .535 win% team right now. If you disagree, change it.

 

All the spreadsheet does is take those 2 skill levels and a league average home/away split to estimate the probability of each game. This does not consider pitching matchups, though. Feel free to adjust the "W PRB" for each game to do so manually. If you think the Brewers have better than a .404 chance of winning tonight with CC, raise it. Think Parra will falter in game 3? Lower the odds for that game.

 

Everytime you hit F9, you run a new simulation of the rest of the season. Watch as the Brewers are both inflated and deflated by the Yost firing dozens of times each!

[davego]

Intersting, according to this we are chasing the Mets rather then holding off the Phillies.

[rluzinski]

If you are just using straight win%, the Mets and Phillies are dead even. They have the same number of wins now and are supposed tto finish:

 

Phillies: .604 x 12 = 7.25 games

Mets: .561 x 13 = 7.29

 

I have to guess that the spreadsheet is a alittle confusing for anyone who didn't make it.

[rluzinski]

I just realized that I don't have the "if needed" Cubs vs. Astros game in there for Houston. Just throw that in the one Astros off day to fix. The probability is going to be way off though, That would most likely be played the day after the regular season?

[dlk9s]

I just did 10 trials using WP3 and we made the playoffs 6 times (Astros weren't close in the games we won, so I didn't add in their final game).

 

I think the lowest win total we had while making the playoffs was 88.

 

 

EDIT: I just started refreshing really fast and hit a scenario with a three-way tie. Neat-o. Time to try to find a four-way tie.

 

EDIT 2: This is fun. I have a sim on the screen right now where we lose our next 5 games and still end up tied with the Mets for the WC. They lose their next 6 games.

[rluzinski]

I simulated it 50,000 with the skill set to win% to get the results on the sim tab. The number on the left are expected win totals (add .4 wins to the Astros). The 38.8% is the percentage for the Brewers winning outright. The others are for tieing with one other team. Didn't look at 3 and 4 way ties. http://forum.brewerfan.net/images/smilies/smile.gif

 

There's a free addon called MCSIM that let's you simulate thousands of seasons very quicly. Easy to use.

 

Yeh, it's a lot of fun to look at some of the crazier results. I like to picture the articles they would inspire. Here's one where the Brewers lose 2 of their next 6 but then win 5 of their last 6, including the win on game 162 to clinch the WC. I should just call this a grit calculator.

[dadofandrew]

If you ignore SL and HOU, the Brewers have to win more games than the lesser of the Mets or Phillies, as one will win the NL East. When you look at it this way, it's easy to think 7-8 wins might win it or tie for it.

[rluzinski]

Only one game behind this pace:

http://i14.photobucket.com/albums/a345/rluzinski/crazyfinish.gif

 

Honestly, I randomly simed that after maybe 50 times. The odds are something like 1 in 1,700.

[BrewCrewRising]

I have a question, but not specific to the excel spreadsheet (Which is really cool, thank you Russ) and I didn't want to create a new thread.

In Baseball Prospectus PECOTA based playoff percentage odds, the Brewers have a higher chance of winning the Wild Card than the Mets, but they average about a game fewer wins in the projections than the Mets. How is that possible?

[dadofandrew]

BCR, that's because the Mets might win their division.

[BrewCrewRising]

BCR, that's because the Mets might win their division.

Except that both the Mets and Phillies finish with higher average wins than the Brewers do. So it stands to reason that since the model predicts that if two teams who are possible WC competitors both finish with higher expected win totals, that those teams should have a higher chance of winning the WC.

 

If the Phillies in the simulation win the division more, it's not like their winning it somehow makes the Mets finish with fewer games than the Brewers and gives the Brewers a better chance at the WC?

 

Not that it matters. After today I'm convinced we're sunk.

[rluzinski]

In Baseball Prospectus PECOTA based playoff percentage odds, the Brewers have a higher chance of winning the Wild Card than the Mets, but they average about a game fewer wins in the projections than the Mets. How is that possible?

 

All the Brewers have to do is finish with more wins than either the Mets OR the Phillies. So, both teams have to finish ahead of the Brewers for the Brewers not to make it. Making up numbers, if the Phillies have a 75% chance of finishing ahead of the Brewers and the Mets have 65%, the odds of them both finishing ahead of the Brewers is:

 

.75 x .65 = 49%

 

That would still give the Brewers a 51% chance of making the playoffs.

 

Of course, with only a few games left, these odds can swing widely. With the Brewers loss today, their odds lowered around 15%. If both the Phillies and Mets wins, we probably are looking at a 1 in 4 shot at the playoffs. Frankly, it's amazing that it's even that high at this point.

[pebadger]

Another way to simulate the playoff race from a Brewer fan perspective:

 

Run full speed into brick wall. Regain consciousness, convince yourself that the result will be diferent next time. Repeat until eliminated.

[rluzinski]

Here's what I get for making the playoffs, depending on the outcome of this series:

 

Current: 36%

1-2: 22%

2-1: 40%

sweep: 60%

 

Of course, that assumes that over the same period, the Mets will win 1.55 games and Phillie, 1.54 games. I'm the kind of guy that tries to root for optimistically realistic stuff, so my goal for the weekend is:

 

Mets: 1:2

Phillies: 1-2

Brewers: 2-1

 

That would raise the Brewers' odds to about 50%. Not going to get worked up over anything bad that might happen though. All done with that.

[SoCal Brew Crew Fan]

Question for you rluz: The playoff odds assume a given probability of a win for each game. They assume little/no correlation from one game to the next. Are streaks more likely in September? Granted, if this is the case, the Brewers are more lilely to continue lsoing versus not...but if they can rattle off a few wins here...is there anything statistically that would show a higher probability of a streak? Just curious.

[rluzinski]

When the result of two games can translate into a 38% swing in playoff odds, it's a big day:

 

http://i37.tinypic.com/au9pow.gif

 

http://i37.tinypic.com/au9pow.gif

 

In about 8 hours, the Brewers could either have a coin flip chance at the playoffs or be a 10:1 longshot. What is the Vegas line on these two games?

 

I assume that each game is it's own even, uneffected by momentum. 99% of momentum is just natural bunching of wins and losses. Very difficult to try and prove that the emotional effects of winning and losing has an impact on performance. Maybe it does but it's very hard to show it.

[zzzmanwitz]

That link isn't working for me russ.

[rluzinski]

Fixed. I should also note that, in the event of a tie, I assumed that the Brewers and Mets had an even chance at winning the tie breaker. Mets get to host, though, correct?

[Oldcity]

Yes the Mets would host.

[fondybrewfan]

According to Accuscore despite the Brewers being down one game they are the favorites to win the wild card. They have the Brewers winning 75% of the time tonight and the Mets winning slightly above 50% of the time (in 1,000 simulations). I am not sure if they take into consideration the array of backups to play for the Cubs in any given game.

 

I would have to think that if the Mets lose and Brewers win tonight, most models would predict the Brewers to win the WC.

[rluzinski]

Today has even a larger potential swing:

http://i33.tinypic.com/29wwkjs.gif

53% swing!

One thing you might notice is that if both teams win or lose, the Brewers odds actually go down. That's because the Brewers have a greater chance of winning today's game, according to my calculations, so they have less to gain with a win.

Another thing to note is that my estimates say that the Mets have the easier final series, since the Cubs are a much better team than Florida , overall. With potentially no Zambrano and Harden and with a many of the starting hitters getting breaks, the Brewers are not even close to facing a full strength Cubs team, though So, keep that into consideration as well.