CSS Button No Image Css3Menu.com
New! Search comments:
(NOTE: Relevance, Author, and Article are not applicable for comment searches)
It has been wrong now for at least 2 months. Time to either fix the description of that column, or fix the formula.
I have the same question as OkayFine.
Blue Jays "average" is higher than all four columns as well. Rockies "average" is lower than all four columns. There are many other examples too.
Even if you were using some sort of weighted average, this would be impossible.
After last year's disaster, it is very encouraging to see how well this report is working this season.
100% chance just means that for all of the simulation runs the Braves made the playoffs. Does not mean they have clinched it.
I would hate to see it discontinued. While it still may not be perfect, it's clear that much progress has been made. We have had to complain loudly, and repeatedly, but our shouts were heard.
Updated this morning at 6:17, but in fact, all the numbers are the same as yesterday.
Some of us noticed that during the All Star break, the playoff odds were swinging wildly every day, even though no games were being played. Are you now saying this was simply due to an insufficient number of simulations?
Very good questions. I wonder if we will get a reasonable answer.
Using the Tigers as an example...
Exp Win%: .527
Sim Win: 83.1
Re-run: 29.2% (I worked it out from the other numbers)
Exp Win%: .527
Sim Win: 83.6
What causes a 10% drop on the re-run? Change in 3rd order wins shouldn't have an effect that big.
It's counter-intuitive, confusing, and misleading. Delta should mean what it says, i.e. the difference between a team's playoff chances yesterday versus today.
Points taken, I'm not sure who is right or wrong. I was already wondering about the issue of actual wins vs. projected wins. But it's still inconsistent. Baltimore and Washington are the two teams in question with more actual wins. And yet they are on opposite sides here. Baltimore has higher playoff percentage than Boston, but Washington is lower than Atlanta.
Note also that BP can't even get the deltas right. If they can't do something as simple as A = B - C, then it's not hard to wonder what else is wrong with this thing.
You could be right, but I'm not convinced. It just feels like too large of a discrepancy. Also the difference in their division chances and wild card chances are in opposite directions. Hard to make a coherent argument for that.
The situation with Nationals and Braves is another example of the same problem.
Apparently if you win 83.9 games in the AL East, you will win the division either 3.9% of the time, or 7.1% of the time, depending which city you play in. And you will win a wild card berth 28.7% of the time, or perhaps 20.9% of the time.
Such are the vague mysteries of the Playoff Odds Report.
Someone from BP please comment on the deltas. What is the point of showing these deltas when they are flat out wrong every single day?
Phillies with a one day delta of 11.0, and they didn't even play yesterday. Really?
Deltas are messed up again today. Tampa goes from 72.4 down to 65.7 and their delta is reported as -1.6? I think it's also fair to ask why Tampa would go -6.7 after beating Boston last night.
What's going on??
Here's a delta comparison from May 24 to May 25. I have only listed the teams that played a game on May 24.
For each team, I show Win or Loss, then playoff odds from May 24 and May 25. Then I show my calculated delta followed by the delta from the report. All of the deltas are incorrect except for San Fran.
Cle W 64.6 65.7 1.1 4.3
Chi W 27.2 29.1 1.9 -0.8
LAA W 39.0 44.1 5.1 4.7
Det L 45.3 40.6 -4.7 -3.5
Min L 0.4 0.0 -0.4 0.0
Sea L 0.4 0.8 0.4 0.0
Cin W 64.2 68.1 3.9 5.5
SD W 0.0 0.0 0.0 -0.8
SF W 37.0 39.4 2.4 2.4
Phi W 26.0 38.2 12.2 9.4
Atl L 81.5 78.7 -2.8 -0.8
NYM L 9.8 5.5 -4.3 -2.8
Mia L 50.4 37.0 -13.4 -11.4
StL L 86.6 87.0 0.4 -4.7
To check the numbers more carefully, I have saved a screen shot of today's playoff odds page. Tomorrow I will compare.
To me, it seems like the numbers are much more sensible than before, just the deltas are broken. For example, Rays playoff pct didn't actually go down by 3.9 compared to prior day, but the delta says it did.
All the numbers are same as yesterday.
Now that this is working, it's interesting to see just how much impact one game can have on a team's playoff chances. Intuitively I never would have expected one day swings of 7 and 8 percent. Very cool.
Yesterday's report was very encouraging, but today kind of feels like the movie Groundhog Day.
The silence from BP is deafening, so let's try this one more time, using Boston as an example.
Boston Red Sox:
"Expected Win Pct" = .534. My understanding is that this is the input to the sim. i.e. this is the chance of Boston winning a game at a neutral site vs. a .500 team. If that assumption is wrong, then obviously the rest of this argument will not make any sense.
"Sim Win" = 86.7, which equals a .535 win percentage. So, at the start of the year, this would be a very sensible result, just a tiny variation from .534 due to luck, and/or strength of schedule.
The problem is, that the Red Sox sit at 12 wins and 19 losses today. So, to get to 86.7 wins, they must win 74.7 of their remaining 131 games, which is a .570 clip. Seems like a long way off from the expected .534, more than we can reasonably explain by luck and schedule.
Repeat this exercise for any team that has started with an unusually high or low win percentage and you will see the same effect.
THE GAMES ALREADY PLAYED ARE BEING IGNORED.
If I seem upset by this, it's because the Playoff Odds Report is a wonderful idea in theory, and I was looking forward to tracking it every day. For baseball nerds like me, it's an extremely interesting way to watch the season unfold. But the execution has fallen far short of my hopes.
Guys at Baseball Prospectus, if you are listening, please give a reply, even if it's to say that I'm completely wrong. Thank you.
So you decide to reply to that minor one-time bug, but not the bigger issue that the actual games played are being completely ignored. Some sort of explanation on that, please? Are today's large 1 day deltas a response to that problem? Or does BP even agree that a problem exists?
If you look at any of the teams that have had an extreme good or bad start, and try to reconcile their final result, it doesn't add up. It's almost as if the sim is ignoring the games already played, and simming the entire season. Look at Washington, San Diego, and the Angels for example. I think some clarification by someone at BP would be helpful.
Angels are currently nine games behind Texas. The sim says that they will finish the season six games behind Texas, even though the input to the sim says the Rangers are the better team. I think we need some clarification from someone at BP that this thing is actually working properly.
Very good point, I hadn't noticed that.
"The expected win percentage used to run the Monte Carlo sim based on projected strength of team, but exclusive of schedule. Not to be confused with actual expected winning percentage."
That's because sometime in the last 7 days they fixed it up to add a 2nd wild card team.
Drabek in the 2012 rotation? That was a joke, right?
I normally just use QERA to evaluate minor league pitchers. I don\'t think this stat would add any value to that, since it includes Hits. Hits don\'t predict anything very well. Add me to the list of those who are surprised that Kevin Goldstein came up with this one.