2014 Strengths of Schedule, Projected
As we know, in any given year, the playing field isn’t exactly even. It’s just one of those things that we quietly accept, because there’s not really much of anything to be done about it. The hope just has to be that, over time, things more or less balance out. (They don’t, at least for teams in the AL East.) For 2014, as for all seasons, the playing field won’t be even. I already took a look at this by projecting divisional WAR. But that can still be taken to the next obvious step — team-by-team projected strengths of schedule, also by WAR.
See, some teams will play easier schedules, overall. Some teams will play more challenging schedules! The effect is relatively small, compared to just levels of talent, but at the extremes it can make a difference of a few wins. Which means it can make a difference between a playoff spot and not a playoff spot, or a protected pick and not a protected pick. From the divisional post, one could already kind of work out the toughest and the easiest schedules, but I thought I might as well just calculate the breakdowns.
I pulled all the schedules off of MLB.com and put them into a spreadsheet. I gathered all the team-by-team projected WARs, as of the moment, and normalized them so that the sum came out to 1,000. For each team, I subsequently calculated the projected average 2014 WAR of their opponents. Here are the results, split by league, with the AL on the left and with the NL on the right, outlined in red for some reason.
The Orioles have the toughest schedule overall, with an average projected opponent WAR of 35.6. The Tigers have the easiest schedule in the American League, with an average projected opponent WAR of 32.9. The Padres have the toughest schedule in the National League, with an average projected opponent WAR of 33.6. The Nationals have the easiest schedule overall, with an average projected opponent WAR of 30.4.
The differences are small. The differences should be small, or else something would be severely broken. But remember, these differences are on a per-game basis, and each team is going to play 162 times, give or take one or two, so there can be a more significant effect. The Orioles will share the AL East with three really good teams and a team that ought to rebound in the Blue Jays. The Tigers will share the AL Central with a couple of borderline pushovers, and, importantly, the Tigers will never have to play against themselves. That’s the part of this that goes kind of unspoken: the worst teams don’t get to play the worst teams, and the best teams don’t have to play the best teams. This right here is one of the reasons a lot of people believe the Tigers have the clearest shot at the playoffs out of the gate.
In the National League, there’s hardly any separation in the middle. Between the fourth- and the 13th-ranking schedules, there’s a projected WAR difference of 1.1. All the way to the right, though, are the Nationals and Braves, who will share the NL East with one tough opponent and three fairly mediocre opponents. Obviously only one of those teams can win the division, but when it comes to determining the Wild Cards, the NL West is deep and for that reason it puts the teams at something of a disadvantage. It’s slight, but so will be the difference in record between the worst playoff team and the best non-playoff team. It’s up to you to decide how much you care about this. It’s a lot easier to just pretend like the schedules basically even out because then you can just concentrate on the baseball.
Yeah, WAR misses some stuff. Yeah, teams are going to change, and there are going to be injuries and transactions, and Kendrys Morales and Stephen Drew and Ervin Santana won’t be free agents forever. Like all projections, these are guesses, and guesses come with error bars and an understanding that they can end up way off. But I don’t think you can really disagree with the conclusions from the graph above. Not today, not given the current MLB landscape. The landscape could change significantly, unpredictably, in the future, but then that’s the exciting part.
Jeff made Lookout Landing a thing, but he does not still write there about the Mariners. He does write here, sometimes about the Mariners, but usually not.
Totally cool graph!
Taking the same data, but added points to show the variation in home/away schedule. The red diamonds show my calculated* advantage for the home/road splits, where positive is better.
http://i.imgur.com/QPT4gLF.png
*How I calculated it
Home/Road split averaged 0.550/0.450 from 2006-2010, so if the game was away the value was adjusted to match.
Then I took that value times 162 to show the effect over a season. A positive value means that if you compare the base WAR-SOS to my aWAR-SOS, the team has an easier schedule.
I would not take those numbers to mean that Baltimore will win 5 more games this year because of an easier home/road split, because I’m not sure how WAR directly correlates to wins. Perhaps Jeff can shed more light on that subject?
I should make virtually no difference what a team’s H/R SOS split is in terms of their overall record.
I’m not saying you’re wrong, but I disagree based on the following: If playing at home gives your team ~20% boost in winning percentage, you would love to schedule based on that.
The Tigers (46) would always prefer to play the Red Sox, Rangers, Rays, Yankees, Angels, Royals, and A’s at home, and everyone else away.
@JunkyMonkey I’m not sure it’s quite like that. Sure, you have a slightly higher chance of beating those teams, but you also have a lower chance of beating worse teams.
If anything, I’d prefer to play more home games against division opponents, rather than good teams. Plus, in the vague “make the fans feel good” way, I’d rather have more overall wins at home than on the road, as long as it doesn’t affect the overall record. (And I think you’d have to prove, statistically, that it affects the overall record before making that claim.)
Travis: It’s true that the gains from playing at home are balanced by the losses from playing on the road. But I think the idea is some teams, by luck of the draw, will get a greater share of games at home where the marginal gains will help them more whilst getting games they’ll still win despite the marginal difficulties of the road.
The Orioles have a schedule 10 times as hard as the Nationals. That is just wrong. Nice graph though.
The schedule is not 10 times harder. Look at the y axis – it starts at 30
10%, not 10x. Big difference.
Look like someone needs to re-read the Y-axis.
Orioles schedule is 17% harder, not 1000%
Um, they do? Eyeballing the chart, it looks like the Orioles are facing teams with an average projected WAR of 35.5, while the Nationals are facing teams with an average projected WAR of 30.3 or so. I’m not a math guy, but that doesn’t seem like a tenfold difference to me. (And keep in mind, that’s not a projected difference of 50 runs per game, it’s 50 runs per year.)
That’s the problem with non-zero minimums on the y-axis, which somebody mentioned in the related post earlier today.
I mean, I don’t wan to be a total fuddy-duddy, but one of the purposes of visual display of quantitative data is to provide a cognitive shortcut to understanding. Non-zero y-axes aren’t unforgivable but they do invite errors in comprehension that aren’t totally the reader’s fault, like here.
I’m just sayin’.
Starting the axis at 0 would make everything look too similar, given the limitations of the page. Maybe that ought to be the point. But what I think I actually should’ve done was noted the non-zero axis in the text, or maybe even on the graph. I’m certainly not pushing an agenda — I just wanted to magnify the area where the differences are. Not a whole lot of purpose in showing the first 30 WAR for every team.
Why not post the graph twice, the first time with the 0-40 y-axis and the second one the 30-36 y-axis you did do.
When I was in grad school in the 1990s, we were taught that the x-axis should be at zero, unless you clearly marked near the bottom of the x-axis (adding a zero below the 30, in this case) with two short parallel lines at a 45 degree angle between the 30 and the 0. This clearly shows that the x-axis is truncated. I agree that this would be optimal.
However, I must add that I’ve never seen this done since getting my Master’s degree. Even Paul Krugman starts many of his x-axes higher than zero. I’m guessing Nate Silver does the same. It just goes to show how low how our once-proud society has sank, and is a sure sign that the anti-Christ will come to Earth soon, if he hasn’t already.
Come on people. Do you have to be spoon fed everything?
Why are people complaining about the non-zero x-axis?
The purpose of the graph is to provide a visual aid to distinguish between these relatively small differences in opponent WAR–not represent opponent WAR to scale and minimize or obscure the differences.
There’s a 30 at the bottom of the x-axis and a 36 at the top. The graph is clear and unambiguous. How brain dead do you have to be to not understand both this graph and that non-zero is the best way to present the graph (irrespective of the most appropriate way to indicate a non-zero graph)?
0 WAR isn’t really all that interesting of a baseline, either. 0 WAR is something like a 50-win team. Starting the graph at 0 projected opponent wins, of course, would make the differences between data points very hard to see.
Maybe an interesting range for this graph would be 40-WAR tall, representing the typical spread between the best and worst teams in the league any given year. Then the graph would show how much schedule strength differs compared to overall talent spread. 6 WAR from top to bottom seems pretty significant, if not enormous.
Hmm, somebody has been reading their Ed Tufte
This site is nerdy enough that we should all know how to read a bar graph. Step 1. Read the main title and axis titles. Step 2. Look at the scale and units of the both axes. Step 3. Look at what’s inside.
So let’s after the season is over we see that the Nationals did in fact play the weakest schedule. Would it then be appropriate to adjust all Nationals players WAR numbers down?
For historical evaluation I don’t see what this adds to simply comparing the Nationals’ opponents wOBA produced/allowed against the Nationals to those opponents’ wOBA produced/allowed against the rest of the league. If you just take aggregate WAR figures, you don’t have any idea how it was distributed; even if the Nationals’ opponents were generally a lousy bunch, they might have saved their best for the Nationals (not necessarily intentionally, just because baseball is random).
Well, it adds the fact that WAR is a much more inclusive stat than wOBA, i.e. park adjusted, adds baserunning and defense, etc.
You can’t do that. That would just be redundant. Whenever a team win against another team the other team tended to play badly against them. If you adjusted each game by the performance of the opposing team, you would just be regressing every game for every team toward the mean. That is not how you want to do SOS adjustments both historically and in the future. It is circular.
Wait so you can’t adjust WAR for SOS? I’m saying adjust outcomes by the true talent level of the opposing team based on how that opposing team did in all other games and how other teams did as well.
Doesn’t Baseball Reference already do this when calculating WAR for pitchers?
Yes, you can do that. I misinterpreted you. You want to use how that opponent team did in ALL games, including against you (adjusting for THEIR SOS of course – it is actually a recursive algorithm).
To the untrained eye, that graph can be very misleading. Good thing all of us are trained.
Taking the same data, but added points to show the variation in home/away schedule. The red diamonds show my calculated advantage for the home/road splits, where positive is better
http://i.imgur.com/QPT4gLF.png
As a Cards fan, I’ll take it!
As a Nats fan, I’m pumped!
As you were last year on March 8th.
As an Astros fan…..**sigh**
Any way that can be spread into opposing teams pitching WAR and Hitting WAR? Would give a sense of what pitching staff is in for a rough year or what offenses may be suppressed a bit.
Jeff, it would have been nice to take one more step and a few extra minutes to translate those differences in WAR to actual win/losses, no?
So does that mean that the Orioles are 1/2 win worse off then the Tigers or 3 wins?
Seems to me it should be a difference of about three wins. Chance I’m missing something obvious.
NO is baaaaaaad
Bad Luck Astros.
http://imgur.com/Tng9i1u
While the Nationals can just march to 1st place.
It’s not bad luck at all, it has a lot to do with the fact that they’re the worst team in baseball. Everyone else gets to play the Astros some number of times, and only the Astros don’t, which is a sizable disadvantage.
But the Orioles are far from the worst team in baseball. According to this graph, the author believes the Blue Jays are a better team than the Orioles. That doesn’t make much sense. But that’s why the games are played.
It is not just the author.
The jays are projected to put up more WAR and more wins than the orioles this year.
http://www.fangraphs.com/depthcharts.aspx?position=Standings
http://www.fangraphs.com/depthcharts.aspx?position=Team
It makes a lot of sense because the Jays are a very good baseball team, and all of the teams in the AL East are pretty darned good.
yup…just like last year’s Nats predictions
the best part of the graph was the acknowledgement at the end that projections are total guesses
i’m excited to see the games actually played
how does DH factor into this equation: Essentially the AL is supposed to have an extra hitter who is capable of starting on their team versus a NL team. Wouldn’t this create a false WAR problem where the extra wins wouldn’t really account for anything as they would cancel each other out except in inter-league play (when 2 al teams play both having a 2 war DH means neither has an advantage)?
If you’ve properly allocated playing time in the projections, then it shouldn’t matter.
can you clarify: Shouldn’t the fact that AL teams have DH’s batting instead of pitchers mean that both players have higher WAR but that higher war will not translate into more wins because that extra non-pitcher will only play against other AL teams (or home inter league games) so given them the extra war from having more non-pitcher innings batting doesn’t accurately reflect the hardship of the schedule .
There are positional adjustments for the DH.
Or you could normalize the data to the Nats 30.4 WAR as 1.0? In this way, the O’s SOS would clearly have a projected ~17% more difficult schedule rather than 1000% (10-fold).
So should i start sipping my champagne with back to back NL West titles? That last time I did that was when Posey hurt his leg……….
To me the difference actually seem quite large, I would think that if your schedule is 10% easier it would lead to 10% more wins. Am I thinking to simple here? If you follow this line of thinking, if the Padres and Nationals would both win 90 games against an equal schedule, than the padres would win 85 games and the nationals 95 games with this type of schedule. Seems like a huge difference to me. Maybe I am forgetting replacement level…
Using the outdated ornithologically correct logo from the days of want and misery makes the task seem harder than it is.
Agreed, use the Cartoon Bird. It looks like every other team has a current logo.
Can someone offer an explanation for 4 out of 5 AL West teams in a row (Angels>Mariners>Athletics>Rangers).
My best guess: the way baseball now schedules games outside the league, each team in the division plays against the same set of interleague foes about the same number of times, but doesn’t play against ALL other divisions in interleague games, just some. The A’s play only the SF Giants and the NL East. I assume that is similar to what the rest of the AL West does- the Rangers play the Rockies and the NL East.
All of them get to play the Astros a lot more than any other team does.
Two years in a row, Orioles have hardest schedule in baseball. This is not fair and should be addressed by MLB.
“Some of the teams are better than others consistently. This is not fair and should be addressed by MLB.”
You have to go back to 1968 to see what getting it right looks like; but even with Divisional play the teams played the same competition the same number of times. It really did make a lot of sense.
Bring back the balanced schedule.
No.
Divisional play is awesome.
Divisional play ruined the game.
We have no idea who the best teams are, if one team gets to beat down the worst team 20 times and another team has to face 4 contenders. We also see teams with weak schedules against the worst teams ousted from the playoffs, while a better team stays home. Its really terrible. I cannot honestly see any reason why anyone would like it, its ruining the authenticity of the game.
Yes. I long for the days when I didn’t know the name of every single Padres player.
I much prefer the balanced schedule to the unbalanced, but the new WC rules mostly negate the problems that SoS differences cause. Sure, SoS might nudge home-field one way or the other, or reault in the 2nd WC being a dif team, but you still only get the 50/50 shot to go into the div round. Much smaller problem than before
I secretly hoped for the Cubs to be in the Nationals position, though I doubt it would make much a difference on their season. But, who knows, I would love a good surprise and they seem to have depth for once everywhere but pitcher.
I have said it before, I’ll say it again – Simply have each team, no matter the league or division, play every other of the 29 teams as close to equal as possible. 29 teams *5 games = 145 games. 162-145= 17 games remaining. So you play every team in your league 6 times, and every team in the opposite league 5 times, and series’ are 5 or 6 games at one shot (except for the last 3 game series), less travel overall, as most series are either 2, 3, or 4 games now. 145+14 = 159. You still have 3 games leftover with a slight edge, 14 more games in your league. So AFTER the 159th game, the final 3 games are scheduled as To Be Determined. They are the WILD CARD GAMES BEFORE THE OFFICIAL WILD CARD. As after the 159th game. you look at all the teams across the entirety of MLB, both leagues, order all 30 teams descending by wins, run differential etc, whatever, so you have a way of breaking all ties, then #1 plays #2, #3 plays 4. #5 plays 6, #7 plays 8 etc, all the way down to #29 vs. 30. Then it could really matter in terms of the top 16 teams, and its a dog fight to get into the playoffs BEFORE the playoffs, in all the close races.
The fans win because they get to see all the other 29 teams on their respective local networks, and baseball is no longer regionalized. As a baseball fan, I would love this. I get to see every team from the comfort of my living room chair. That way, the schedule would be as even as it gets too.
Right now the MLB players make about 47-55 flights during the regular season. Especially given the 2 games “series” they play during interleague. The way I suggest would cut cost of travel, players would not be worn out due to travel as much as the way I suggest would be 29-32 flights only, considering possible rainouts.
Oops. my math is a bit wrong in the first comment. Say you are an American League team, so you play 6 games * 14 other teams in the AL (not counting yourself) = 84 games. 15 teams in the NL * 5 games = 75 games. 84+75 = 159, + 3 to be determined wild card, because the team you are actually playing depends upon standings and is an unknown for most of the season. So 84-75 is only 9 games more in the AL, not 14 BEFORE the last 3 games of the year, where you could theoretically play a team in either league. So Only 12 games more played in your teams league at most.
The problem with each team playing all their games against each opponent at once is that–to be fair–each team would need 3 games at their home ballpark (based off your 6 game per team (initially) proposal). Wouldn’t that put the travel back to where it is?
Divisional play is awesome.
Especially since I have no interest watching the Astros/Twins play.
Then don’t watch. I’d venture to guess that you aren’t currently watching, so what’s the difference?
There is likely a hybrid solution available, which preserves divisional play (to promote rivalries, increase chances of a fan’s games being within their own timezone, etc.) yet also tries to make the rest of the schedule balanced. Now that the divisions are all even it could be easier to accomplish.
For example, 25 non-division rivals to visit, 10 in one’s own league, 15 in the other… 3 home and 3 away in one’s own league is 60 games. 2 home and 2 away in the other league is 40. Visit the 4 teams in your own division 15 times (8 home, 7 away), that’s 60. No, it doesn’t get to the hallowed 162 game number, but it keeps divisions reasonably important while getting as close to balanced otherwise as possible. Perhaps those extra 2 days, plus shortening spring training, could help make it so that the World Series is likely to end before Halloween.
I’ve maintained for years that if one could accommodate or simply remove some of the club preferences (e.g. Nats always want to play at home on July 4), one could arrive at a One True Schedule rather than redoing it every year.
I know the major hurdle here is asking the 30 owners to lose 2 games worth of revenue. I don’t have an answer for that.
Yes, there is something to be done about it:
Step 1: do away with the concept of divisions, just have 2 leagues, and the top 4 or 6 teams make the playoffs.
Step 2: Have each team play every other team in their league 12 times (6 home, 6 away). 12 X 14 = 168 games played.
This model is simple and effective: we will finally see who is good and who is just taking advantage of having 20 games against the worst team in the league.
If you must have playoffs. What is determined in playoffs that isn’t settled in a 168 game schedule? The league champions play in the WS. End of story.
One thing that’s missed here is that WAR is partially derived from last year’s schedule, because player’s last year stats are created against that schedule…
So there’s really a second order adjustment that needs to be made. Because WAR doesn’t have any quality of opponent/SOS adjustment, guys who play tougher schedules will have their WAR suppressed.
So the real SOS differences are probably larger than this shows.
I didn’t bother reading the article because I have spent the last ten minutes staring at that totally pretty graph, and just wanted to jump down to comment how amazingly beautiful it is.
Wow!
Thanks for posting this, and I hope to read the article some time this week. Thanks!
This is a pattern in baseball. Baseball gives small advantages over the course of 162 games to large market teams. I would not call game fixing but rather strategic management of wins to ensure that baseball as a whole increases market penetration and profits. Boston dumps quarter of a billion in dollars in bad contract to the Dodgers. This was reckless. AL east team suffer from return of the good Soxs. Incidentally, LA get one of the easiest schedules in the majors. Oh before I forget NY got the ARod salery relief program. I wish some would do a 5 study.
Looks like you put a lot of work in to this. The schedule is very helpful.