Football Outsiders vs. Vegas: 2005-2009

Back in January, I posted a comparison of how the Football Outsiders win projections have done against the Vegas line over the past few years. With the season about to start, I thought I'd take another look at this comparison, with a few improvements.

First, with the help of a couple readers/Twitter followers, I was able to find Pinnacle's win totals from September 2005. I had to go back a week or two for a couple that were OTB for whatever reason, but the majority of those totals are from fairly close to Week 1.

I'm also now using the final win projections posted on the FO website (2009, 2008, 2007, 2006, 2005), whereas I was previously using the numbers published in their book. Since they have a lot more information a couple days before the season than when the book goes to press, this should help FO.

Finally, as I do with similar baseball analysis, I'm using RMSE rather than average miss.

Here are the standings for the last five seasons:

No matter how you look at it, the FO projections did not fare very well in 2009. Schatz has an interesting post from a couple months ago talking about what happened, if not why.

But while that is the most recent example, it's not the end of the world. The margin was almost as large in 2007, only then it was FO with the significantly lower number. Over the last three years the two "systems" are tied, and since 2005 the Vegas lines are only slightly ahead.

Another way to look at this is throw the two sets of predictions into a regression and see which one is weighted more heavily. Doing that for all five years gives the equation of:

Actual wins = 1.42 + 0.44 * Vegas + 0.39 * FO

This looks a lot different than what I got in January, which is not surprising because of the changes/additions described above, but I didn't expect the weights to be so close. When FO posts their final numbers next week, I'll have another post looking at how the numbers compare, and where they really differ.

This Week's Links (8/23-8/27)

Still trying to track down two of the five winners from last year's Survivor pool -- Shnux, and scoot2929. You can still sign up for this year's group as well, there will be some sort of small cash prize.

RLYW with what I think is a great analysis of the AL Cy Young race.

Tango's Fans Scouting Report is back.

Football Outsiders' Bill Barnwell answers questions from Freakonomics readers.

Luck, timing, and win distributions in baseball.

2010 Totals Contest Update: August 23

Exactly three months after the last update; I totally planned it that way. I calculated expected points the same way I did last time, by estimating the % chance each team has of going over/under their total. The spread is a bit higher this time, as the Padres are already over their # of 71.5, and Seattle would have to go 35-3 to go over 83.5.

Here is the updated top 10:

Things are going very well for "wadd", as seven of his 10 picks (ATL O, SFG O, MIN O, BAL U, CIN O, PIT U, SDP O) are 80+% to win, and his other three are about 50-50. If Oakland and Houston make it over their totals, he will likely be our winner.

Everybody's projected points and place in the standings can be found in a Google Doc here.

Edit: Hoping to get back into a more regular posting schedule in the near future, and I'll definitely be doing the weekly NFL Survivor posts again. On that note, I've created an ESPN group. Also, if you're one of the five people who went 17-0 in our pool last year, please email me, as I need to set up the "winners pool" I promised.

This Week's Links (7/5-7/9)

Sorry for the long delay in posting. Hopefully someone posts MVP/Cy Young odds over the All-Star Break so I can analyze those next week.

Good stuff from Baseball Analysts on bimodal pitch locations.

Does the ball, indeed, not lie?

Ruben Amaro Jr. trying to explain how trading Cliff Lee in December for an uninspiring package of prospects, and now looking to use prospects to acquire pitching help at the deadline, makes sense is as entertaining as you'd expect.

MGoBlog on David Berri.

KenPom brings us the ten unlikely wins of the 09-10 college hoops season (10-6, 5-1).

Interesting analysis from Dave Allen on MLB's Postseason Ticket Reservation program.

It's hard to like some NBA superstars these days. Kevin Durant does not fall into that category.

Big 12 basketball is going to be quite strong next year.

Updated Strasburg Odds

It never ends, this time at Pinnacle and Sportsbook:

Total 2010 Wins (Sportsbook)
Over 9.5, -120
Under 9.5, -120

Total 2010 Wins (Pinnacle)
Over 7.5, -120
Under 7.5, +103

Well, he has one under his belt and is now at 62.3 innings on the year. He went seven innings in his first start, but those were somewhat unusual circumstances; I'll say he averages 6.5 innings per start the rest of the way. Conveniently, that puts him right at 15 starts the rest of the season if he can only pitch 160 innings.

Pinnacle's total doesn't seem to have been effected too much by his performance on Tuesday, beyond the fact that he's now one win closer to 7.5; their juice was +153/-180 on Tuesday afternoon.

Another consideration here would have to be his opponents, which set up quite favorably:

6/13: at Cleveland
6/18: vs. Chicago (AL)
6/23: vs. Kansas City
6/28: at Atlanta
7/3: vs. New York (NL)
7/8: vs. San Diego

Four home games out of six (making five out of seven total), and that's not exactly a murderer's row of offenses. I'd expect the Nats to be favored in five of those six games, and probably underdogs against Atlanta depending on the opposing pitcher.

Winning more than 9.5 games would involve getting a W in nine of his 15 remaining starts (60%), while the percentage needed to go over the Pinnacle total is 47%. I would definitely take the under on the Sportsbook total, while the Pinnacle number seems about right, if slightly low considering the opponents he'll be lined up against.

2010 ERA (Sportsbook)
Over 2.75, -130
Under 2.75, -110

2010 ERA (Pinnacle)
Over 3.00, +120
Under 3.00, -140

This is where I have to jump off the bandwagon. These are nuts. According to CHONE, here are the six lowest neutralized ERA projections for SPs:

I was there on Tuesday night. I know he's good. Really good. But forget 2.75, projecting a pitcher to have an ERA under 3.00 is extremely rare. An interesting poll conducted by Tangotiger (beginning on Wednesday morning) agrees, pegging Strasburg for a 3.36 ERA the rest of the season. I'm not sure if O2.75 -130 or O3.00 +120 would be the right choice, but I think that's the right side here.

2010 Strikeouts
Over 146.5, -120
Under 146.5, -120

It may be that Sportsbook is assuming Strasburg will make regular starts the rest of the season, rather than stopping once he gets to 160 innings, because this one is psychotic. Fourteen Ks in seven innings is a good start, but that still leaves him needing 133 more in his 97 2/3 remaining IP, which works out to 12.3 K/9. Even during his dominant minor league run, Strasburg "only" struck out 10.6 batters per nine innings; the under is the clear choice here.

Will Stephen Strasburg throw a no-hitter in 2010?
Yes, +550
No, -1000

This is an interesting one. What we really need for this is his expected batting average against, which you can basically calculate from K%, HR%, and BABiP. I'm going to be much more generous than his minor league translations suggest, and go with a 10.0 K/9, which works out to about a 27% K%. We'll also go with 0.5 HR/9 (1.37%), and a .285 BABiP (again, generous). That works out to a BAA of .218, giving him a 0.13% chance of pitching a no-hitter on any given night (1 in 761), and a 2.0% chance of throwing at least one in his final 15 starts.

All of that is assuming they'd ignore pitch count considerations and keep him in to finish off a no-hitter, which they likely would but it's not a sure thing. Regardless, +550 is clearly a terrible bet here, and "No" is actually very strong.

More Strasburg Odds

Pinnacle released a bunch more Strasburg odds today, including the simplest yet possibly most interesting line:

That's what Pinnacle is offering as I write this, but it's actually nowhere close to the line they initially released. They originally offered Nats -164/+154 on Monday afternoon, but that got bet up very quickly.

The Nationals have not been favored this heavily in nearly five years. From Covers.com, here are the ten most lopsided spreads in Washington's favor since they moved to the nation's capital:

Those top two, on back-to-back days in July 2005, came when the Nats sat atop the NL East, while the Rockies had barely won 1/3 of their games.

Another thing I was curious about was the debut lines for other highly touted minor leaguers. Here is a fairly arbitrary selection of eight over the past decade (Covers' database only goes back to '99):

Not that you can make direct comparisons between pitchers facing the same team at home eight years apart, but the Prior line is interesting. This is obviously a very favorable matchup for Strasburg, and it's possible that some NYY or BOS pitcher was a huge favorite at home in his debut against the Royals or something, but being -200 in your major league debut is pretty crazy, and maybe unprecedented.

Now onto some of the other Strasburg props at Pinnacle:

Total Wins in 2010
Over 7.5, +153
Under 7.5 -180

In yesterday's post I thought there was value in U8, and clearly Pinnacle agrees with me there. This line seems much more reasonable.

2010 ERA
Over 3.50, +112
Under 3.50, -131

Interesting, if not too surprising, that Pinnacle thinks he'll win fewer games than the SBET line indicated, but with a lower ERA. This line, coupled with the side, proves to me that all the hype is completely justified.

I'm not going to go through these individually, but Pinnacle is also offering props on how Strasburg fares on Tuesday night:

Strasburg's Personal Result:
Win, -139
Loss, +251
ND, +247

Total Strikeouts
O5.5, -156
U5.5, +133

Total Pitches
O90.5, +129
U90.5, -151

Total Innings Pitched
O6, +110
U6, -129

Will He Allow A HR?
Yes, +169
No, -200

Earned Runs Allowed
O2.5, +117
U2.5, -137

So he's expected to strike out over 5.5 guys, while pitching under six innings. Not bad, for a rookie.

Stephen Strasburg Odds

As you may have heard, Stephen Strasburg will be making his major league debut for the Nationals on Tuesday night against the Pirates. SportsBetting.com posted some interesting props on how successful Strasburg will be in his rookie campaign, a few of which I break down below. I might do another quick post on this subject after Pinnacle releases a line for the game on Monday evening.

Total Wins in 2010
Over 8, -125
Under 8, -125

Between AA and AAA, Strasburg has thrown 55.1 innings so far in 2010. The Nationals apparently are targeting about a 20% increase from last year's innings total, which would put him at 160.

From that point, the number of starts he ends up making depends on how many innings he averages per start. He's been at exactly five so far this year, but I suspect the Nationals will let him go a little longer than that now that he's with the big club.

That works out to about 17 starts the rest of the year at six innings a pop before he hits his innings limit. Notching a win in eight of those games would mean he records a victory in 47% of his starts. From 2005-2009, there were 453 pitcher seasons in which the pitcher started at least 24 games and didn't make any relief appearances; only 91 of those won at least 47% of their starts (20% of the population).

Even if we limit the population to pitchers with an ERA+ of 120 or better, only 54 of the 131 eligible pitchers won 47%+ of their starts (41%). Strasburg is very good, but a 120 ERA+ isn't bad either (and that was the minimum in that group). Those pitchers also averaged 6.55 innings per start, and likely played on better teams than the Nationals. Throw in the possibility of injury, and I think the under has a clear edge here.

2010 ERA
Over 3.65, -120
Under 3.65, -130

This is tough. Normally I'd look to projections for something like this, but that's not going to be nearly as effective in this particular case. Sean Smith released updated CHONE projections last week, which put Strasburg's ERA at 4.37. However, Sean immediately notes that this is a very pessimistic projection, since it only knows about Strasburg's 2010 statistical resume, and not his college dominance, draft position, fastball velocity or nasty breaking ball.

I find myself wondering how SBET came up with this line in the first place; I'm assuming it was an inexact science. Bodog released a line on the same prop about three weeks ago with the O/U being 3.5; with the under being favored on SBET's prop, that is pretty similar. Looking at the updated CHONE projections, a 3.60 ERA would put him 12th among starting pitchers in nERA (I'm assuming that's a neutralized ERA), tied with John Danks and Josh Johnson. A high bar, certainly, but I wouldn't want to take a side on this one either way.

Will Stephen Strasburg win the National League Rookie of the Year award?
Yes, +775
No, -2000

While looking into the two props above I have essentially set Strasburg's expected line for 2010 at 7-3 with a 3.60 ERA. When handicapping his RoY chances that's not all that helpful, since the error bars are so high, but it's a starting point.

Strasburg probably didn't pick the best year to try to win this thing after being called up two months into the season. The hype helps, but there's a certain outfielder in Atlanta who already has 219 major league PAs and has hit .274/.402/.531 with 10 homers while receiving a bit of attention himself.

Even beyond Heyward, there's going to be some serious competition for best rookie pitcher in the NL, as the slightly less hyped Mike Leake is 5-0 with a 2.22 ERA, and a 3.51 FIP. I doubt this will be in the forefront of the voters' minds, but the Cincinnati rookie has also been tremendous at the plate, with a .440 OBP in 29 PAs.

Finally there's Jaime Garcia, who has been just as good as Leake, at least on the mound, with a 1.32 ERA and 3.02 FIP in 61.1 innings going into Sunday night's start. There's chance that Strasburg goes nuts and ends up something like 10-1 with a 2.00 ERA and a ton of Ks, but that can't be all that likely, and I think the line SBET posted on this is pretty sharp.