Spycake, thanks for passing along that link to the FanGraphs. I love that they spell out the probabilities for different methods. It clarifies a lot about these.

 

1. Even with the Twins in 2nd place for the Wild Card right now, they have the worst chance to get it among the contenders using the Fangraphs Projection Mode. That's because it's based on the projections of the players that FanGraphs made (which had the Twins finishing with the worst record in the AL in the offseason). I don't know how often they updates those projections, but it looks like it makes a pretty big difference because...

 

2. The Twins odds go from 8.7% to 21.3% on FanGraphs "Season To Date Stats Mode."  This uses this year's stats: 

This mode uses current season stats, weighted more heavily towards the most recent games, to calculate the winning percentage of each remaining game in the major league season.

 

3. And those odds skyrocket up to 24.1% in "Coin Flip Mode" which might be the most useful, and is certainly the most transparent. The Twins odds rival that of the Rangers (25.3%) and are higher than the rest of the field. 

 

I prefer my own back-of-the-napkin mode: There are basically five teams bunched together for the last wild card spot. It sounds to me like they each have about a 20% chance of winning it. 

That sounds like a good summary, John. Although I will note that your back-of-napkin method becomes a lot less elegant as the competitors get more spaced out. If there is a leader and one team 2 GB and another 4 GB, I really wouldn't know how to assign percentage odds to them all. Also, at 40 games out it doesn't matter as much, but a 2 game lead will become increasingly significant, and it is hard to tell exactly how much with just an eyeball glance at the standings.

 

 

I prefer my own back-of-the-napkin mode: There are basically five teams bunched together for the last wild card spot. It sounds to me like they each have about a 20% chance of winning it. 

Lol … that's kind of my thinking.

 

I prefer my own back-of-the-napkin mode: There are basically five teams bunched together for the last wild card spot. It sounds to me like they each have about a 20% chance of winning it. 

Plus or minus 19%.

 

That sounds like a good summary, John. Although I will note that your back-of-napkin method becomes a lot less elegant as the competitors get more spaced out. If there is a leader and one team 2 GB and another 4 GB, I really wouldn't know how to assign percentage odds to them all. Also, at 40 games out it doesn't matter as much, but a 2 game lead will become increasingly significant, and it is hard to tell exactly how much with just an eyeball glance at the standings.

 

You're right. And as we get closer to the end of the season, those gaps mean even a bit more. That, by the way, is when the percentages can really jump around a lot, too. And they should. 

 

But right now, there isn't much of a gap. Let's hope this 4.5 months of decent baseball ends up giving us a few weeks of some nightly drama and lost of mornings of overreaction. 

Now the mlb.com site is giving the Twins a 14.4% chance, up significantly with the wins this weekend and last night. Tampa, despite being two games behind the Twins has a higher probability, but the Orioles have fallen behind. Angels and Rangers also have a higher projection than Minnesota.

Fangraphs has the Rays behind us too.

 

In their season-to-date stats mode, we actually have the best chance at the second wild card!  And an even better chance in the coin flip mode.

Yeah, BP's model seems excessively down on the Twins. I'm curious which part of their model is causing that drag.

 

Yeah, BP's model seems excessively down on the Twins. I'm curious which part of their model is causing that drag.

 

Run differential, maybe?    The Twins have trailed the other wildcard contenders consistently in that stat, and even with the recent hot streak they still trail three other WC teams

Run differential, maybe? The Twins have trailed the other wildcard contenders consistently in that stat, and even with the recent hot streak they still trail three other WC teams

Even so, run diff shouldn't weigh that heavily into the results.

 

Even so, run diff shouldn't weigh that heavily into the results.

Agreed.  I was just guessing at their methodology, not defending its validity.

 

I was hesitant to even comment without looking at the numbers more closely, but RD has been a concern pretty much all season.

 

And yes, its value as a predictive tool declines after trade deadline moves and roster expansion, and with the decreasing sample size of games remaining.

 

The battle for the second AL wilcard spot is shaping up to be a knife fight in a phone booth, and matchups have more to say about the outcome than aggregate stats.

Yeah, BP's model seems excessively down on the Twins. I'm curious which part of their model is causing that drag.

It's almost certainly their player projections. BP and Fangraphs both project the Twins to be the worst team in the AL the rest of the way (and have been doing so for some time ), and their projected odds are a near match.

 

Taking out the player projections at Fangraphs reveals much better odds for the Twins.