clock menu more-arrow no yes

Filed under:

How many games will the Houston Astros win in 2015?

New, 41 comments

Trying to project how many games the Houston Astros will win next season, using a range of projection tools.

Troy Taormina-USA TODAY Sports

With the new season fast approaching one question is on everybody's mind: how many games will the Houston Astros win in 2015? Depending on who you ask the answer will differ. Some will tell you they have no chance whatsoever and will wind up losing a lot games. Others, the more optimistic among us, project a five hundred season or something close. And, finally, the dreamers think that maybe, just maybe -- with a couple of lucky breaks -- the Astros can sneak into the wild card game. I'm attempting to answer that question and shed some light onto what the Astros can do next season.

Naturally, there are several ways to try and project how many games the Astros will win next year. However, that said, I'll be using just two projection methods: the Pythagorean win formula -- using runs scored, and runs allowed -- and also, quite simply, adding up the total projected WAR of the roster, from several different projection sources. So, without further ado, I'll begin with the WAR method.

The first thing to remember is these are wins above replacement. Which, as always, begs the question: how many games do a replacement level team win? Well, quite simply, 48 games. For a little additional information on how that figure was magically conjured up, here's what FanGraphs have to say on the matter. Hence, rather simply, the total number of projected wins the Astros tally is added on to 48 to give us how many games the Astros will win next season.

This new unified replacement level is now set at 1,000 WAR per 2,430 Major League games, which is the number of wins available in a 162 game season played by 30 teams. Or, an easier way to put it is that our new replacement level is now equal to a .294 winning percentage, which works out to 47.7 wins over a full season.

To begin, let's look at the most likely 25-man roster as outlined by MLB.com and with some help of the TCB staff: Chris Carter, Jason Castro, Hank Conger, Jon Singleton, Jose Altuve, Jed Lowrie, Matt Dominguez, Jonathan Villar, Marwin Gonzalez, Jake Marisnick, Dexter Fowler, Robbie Grossman and George Springer are the hitters. And the pitching staff will likely look something like this: Scott Feldman, Dallas Keuchel, Collin McHugh, Brett Oberholtzer, Samuel Deduno, Chad Qualls, Pat Neshek, Luke Gregerson, Tony Sipp, Josh Fields, Mike Foltynewicz and Will Harris.

Starting off with ZiPS projections, done by Dan Szymborski, here is the projected WAR for every player that will likely stay on the 25-man roster for the full season -- with help from the aforementioned MLB.com depth charts and TCB staff. The 25-man roster is a rough estimate, please remember. I'll calculate the totals after we've looked at the Steamer projections, too, which are in the following table. Quick note: you should probably brace yourself for bad projections, sadly.

Player (ZiPS) pWAR
Hank Conger 1.2
Jason Castro 2.7
Jon Singleton 0.5
Jose Altuve 3.6
Jed Lowrie 1.6
Matt Dominguez 0.8
Marwin Gonzalez 0.7
Jonathan Villar 0.6
Chris Carter 1.7
Dexter Fowler 1.8
Jake Marisnick 1.2
George Springer 2.7
Robbie Grossman 0.0
Scott Feldman 1.0
Dallas Keuchel 3.0
Collin McHugh 2.7
Brett Oberholtzer 0.4
Samuel Deduno 0.2
Chad Qualls -0.1
Pat Neshek 0.4
Luke Gregerson 0.5
Tony Sipp 0.3
Mike Foltynewicz -0.9
Will Harris 0.4

Player (Steamer) pWAR
Hank Conger 0.6
Jason Castro 1.6
Jon Singleton 0.6
Jose Altuve 3.0
Jed Lowrie 2.1
Matt Dominguez 0.9
Marwin Gonzalez 0.1
Jonathan Villar 0.0
Chris Carter 1.5
Dexter Fowler 1.7
Jake Marisnick 0.1
George Springer 2.5
Robbie Grossman 0.3
Scott Feldman 0.9
Dallas Keuchel 2.7
Collin McHugh 2.1
Brett Oberholtzer 0.6
Samuel Deduno 0.2
Chad Qualls 0.4
Pat Neshek 0.3
Luke Gregerson 0.6
Tony Sipp 0.3
Mike Foltynewicz -0.1
Will Harris 0.0

While there will be, more likely than not, plenty other players that will contribute over the duration of the season -- especially in September -- to keep things simple I'm just sticking to the 25-man roster. Besides, the surplus of players who may potentially contribute later in the season mainly own negative WARs. Now, let the calculations begin. Naturally, I'm just going to total the two projections and find the average.

ZiPS Steamer Average
27.0 23.0 25

According to the average of the Steamer and ZiPS projection systems, the Houston Astros are worth 25 wins above replacement. And, now, in order to complete the WAR calculation, all we must do is add the 25 wins above replacement the Astros are worth to the 48 games a replacement level team would win.

RL pWAR Total
48 25 73

Well, there you have it. According to ZiPS and Steamer projection systems the 2015 Houston Astros will win 73 games. 73 wins is hardly devastating, but still way off how many games I think the Astros could and should win next season -- which, for the record, is around the 81 win mark. But, obviously, WAR projections aren't the only way to attempt to calculate how many games the Astros will win. So, now onto the aforementioned Pythagorean win formula.

Going by the major league average from last season, I'm going to presume the Astros will amass 6131 PAs and pitch 1458 innings -- or, 162 nine inning games. Calculating how many runs the Astros will allow could be very easy. All that has to be done is divide the Astros' projected ERA by the amount of innings they're likely to pitch. However, I'm going to calculate how many runs each individual pitcher will yield and then scale it to the league average for innings pitched from last season.

Then, I'll calculate how many runs the Astros are projected to create and work out how many they'll create if they amass the league average 6131 PAs from last season and fill it into the pythagorean win formula. Let the fun begin. I'll start with the pitching and once more, for ERA, I'll take Steamer and ZiPS' projections and find the average for each pitcher on the aforementioned 25-man roster. While, yes, there are more advanced metrics available, I'll stick to ERA. It's easy, and accurate, providing the projection is, in calculating how many runs each pitcher will allow.

Pitcher ZiPS Steamer Average
Scott Feldman 4.48 4.91 4.70
Dallas Keuchel 3.80 3.82 3.81
Collin McHugh 3.71 4.09 3.90
Brett Oberholtzer 4.82 4.70 4.76
Samuel Deduno 4.79 4.66 4.73
Chad Qualls 4.25 3.71 3.98
Pat Neshek 3.51 3.47 3.49
Luke Gregerson 3.52 3.67 3.60
Tony Sipp 3.69 3.13 3.41
Josh Fields 3.74 3.16 3.44
Will Harris 3.74 3.78 3.76
Mike Foltynewicz 5.53 4.09 4.81

Onto how many runs each pitcher will allow, based upon how many innings they're likely to pitch --- as projected by just Steamer, but it will be put onto a scale of 1458 innings pitched, anyway -- and their average ERA from the two different projection systems.

Pitcher pERA pIP pRA
Scott Feldman 4.70 191 100
Dallas Keuchel 3.81 188 80
Collin McHugh 3.90 188 81
Brett Oberholtzer 4.76 150 80
Samuel Deduno 4.73 48 25
Chad Qualls 3.98 55 24
Pat Neshek 3.49 65 25
Luke Gregerson 3.60 65 26
Tony Sipp 3.41 55 21
Josh Fields 3.44 40 15
Will Harris 3.76 30 13
Mike Foltynewicz 4.81 45 24

Now, the calculations. I'll break each stage down into a table to make it easier to follow and to make it easier for me to keep track of. Another, quick reminder: my maths abilities are about as good as Matt Dominguez is at making contact. He's still an everyday third baseman, and I'm still going to try and work this all out.

Total Innings Pitched Total Runs Allowed
1120 514

With the totals all added up, the more difficult calculations -- to the simple minds like mine -- can begin. I'm going to work out both runs per inning, and runs per game. After all, better safe than sorry. Both calculations, as usual, will be put into the table below.

pRA/IP pRA/G pRA/1458 pRA/162
0.45 4.05 656 656

Thankfully, the totals are the same, and I'll continue to trust my maths ability for the duration of this article, you'll be glad to hear. Essentially, to infer the results into something meaningful, the Astros will allow 4.05 runs per game next season, and will allow 656 runs on the year -- according to ZiPS and Steamer projections. Now, moving onto the offense. How many runs will the Houston Astros offense score next year? I'm going to use Steamer projections for PA and wRC+. The first thing to work out, before looking into the Astros, is how many runs did an average player create last season. The constants are supplied by Fangraphs. (The average PA/G for the AL last season was 3.28)

R/PA R/G R/162
.108 .354 57

So, what we can infer from this data is that the average player created .354 runs per game, and in turn, the average player created 57 runs last season. Now, by looking at each players' wRC+, and the percentage of runs they created above/below average, the total number of runs each player is projected to be worth can be calculated, relative to the league average. Once that has been done, I'll work out how many runs the Astros will create next season, once more, relative to the league average.

Player pwRC+ pPA RC/PA tRC
Jason Castro 92 464 .099 46
Hank Conger 86 282 .093 26
Jon Singleton 104 523 .112 59
Jose Altuve 113 688 .122 84
Jed Lowrie 105 525 .113 59
Matt Dominguez 85 606 .091 55
Dexter Fowler 109 526 .118 62
Jake Marisnick 78 562 .084 47
George Springer 118 493 .127 63
Marwin Gonzalez 78 361 .081 29
Chris Carter 116 553 .125 69
Robbie Grossman 95 604 .103 62
Jonathan Villar 78 527 .084 44

And there you have it, the projected runs created for each player on our hypothetical 25-man roster. Now, to try and calculate how many runs the Houston Astros will score, based upon the 6131 PAs the average team amassed last season. I'll simply work out how many runs the Astros are projected to create per plate appearance, and in turn, how many runs they will create over 162 games.

tPA tRC RC/PA RC/162
6714 705 .105 644

So, according to my somewhat accurate calculations, presuming the Astros amass 6131 plate appearances and pitch 1458 innings, the Astros will score 644 runs, and allow 656 runs. If I fill that into the Pythagorean win formula, I'll know how many games the Astros are projected to win next season. As usual, I'll fill the results into the following table.

Runs Scored Runs Allowed Games Played Games Won
644 656 162 80

According to the Pythagorean win formula, the 2015 Houston Astros will win 80 games, and hence finish the season off at one game below five hundred. And there you have it: the Pythagorean win formula thinks the Astros will finish the season 80-82, while a combination of Steamer and ZiPS projected WAR project the Astros to finish 73-89. Personally, I'm siding with Bill James and the Pythagorean win formula, this Astros team has the potential to finish the season around the five hundred mark.

All in all, I'd expect the Houston Astros to win around 81 games. I think the WAR projections are a bit harsh, and the Pythagorean win formula pretty accurate. So, who will be right at the end of the day: the aforementioned pessimists, who predict another bad season, the optimists who predict a five hundred season, or the dreamers, who think we can sneak into the playoffs? Only time will tell.