Does It Matter Who You Beat On the Way to the Playoffs?
The Mets have probably had the easiest schedule in baseball. That’s not to take anything away from them — it’s not their fault, and the Nationals have faced pretty much the same slate. It’s just a fact. Things have conspired to make the Mets’ schedule fairly soft. One notices, if you dig in, the Mets have really taken advantage of this. Against teams under .500, they’ve won 67% of their games. Against teams at least .500, they’ve won just 41% of their games. Against the Phillies, the Mets have gone 14-2; against the Pirates and Cubs, they’re 0-13. Their difference in record by opponent quality is the biggest in the game, edging out the Orioles and the Dodgers.
At the other end of the extreme, you find the Blue Jays. The Mets have mostly beaten up on bad teams. The Blue Jays have beaten everyone, but especially the above-average teams. Against teams under .500, they’ve won 53% of their games. Against teams at least .500, they’ve won 63% of their games. Their difference in record by opponent quality is the biggest in the game, in the other direction. They beat out the A’s and the Tigers.
With the playoffs looming, it’s easy to speculate. Once October rolls around, only good teams are left. It seems like the team that’s been better against good teams should stand a better chance. Does this mean anything for the Blue Jays and Mets? Below, I can provide not necessarily the answer, but certainly an answer.
I want to be straightforward with you: there are probably better ways to do this. There are certainly better ways to do this. This is going to be a little rough, but it’s also the way that the Baseball-Reference Play Index made really easy. So we’ll manage. Obviously, not all .500+ baseball teams are created alike, but an assumption here is that that sort of noise evens out. Ditto the teams under .500. You can take more liberties when your sample is decently large.
I decided to cover the newest era, the wild-card era, dating back to 1995. For every playoff team, I collected winning percentage against good teams, and winning percentage against bad teams. Then I folded in how those teams actually went on to do in the playoffs. It should be obvious there are a lot of variables here, messing with things, but it stands to reason if there’s any sort of strong effect, it ought to show up. Here’s a table. There have been 166 playoff teams, so I split them into four groups, two of 41 and two of 42. They’re grouped by their performance differences against good and bad teams. Group 1 was the strongest against good teams; Group 4, comparatively, beat up on the bad teams.
The columns should be easy enough to understand. Difference = win% vs. .500+ opponents – win% vs. sub-.500 opponents

| Group | Season Win% | Difference | Playoff Wins | Playoff Games | Playoff Win% |
| Group 1 | 0.581 | -0.009 | 150 | 309 | 0.485 |
| Group 2 | 0.573 | -0.085 | 168 | 338 | 0.497 |
| Group 3 | 0.587 | -0.133 | 185 | 360 | 0.514 |
| Group 4 | 0.585 | -0.210 | 154 | 307 | 0.502 |
I wasn’t sure what to expect. Maybe you weren’t, either! And even here, for each group, we have playoff samples between 307 – 360 games, so it’d be super to have more data, but if this table is any indication, there’s no benefit from having beaten more quality opponents. Group 1 had the smallest difference between success against good teams and success against bad teams. It also shows the lowest playoff win percentage. Group 4, meanwhile, beat up on worse teams, relatively speaking, and that didn’t hurt them. Group 3 shows the most playoff success, but Group 3 also includes the best regular-season teams, as shown in the second column.
The 1999 Yankees did very well against good teams, and they won the World Series. But then you have the 2002 Angels — they won the World Series, too, after losing more than half their games during the season against .500+ opponents. They won three-quarters of their games against sub-.500 teams. The 2010 Giants won the World Series with a similar profile to the 2002 Angels. Of the four teams with the strongest performances against good opponents, compared to bad opponents, they won just two of 14 playoff games. It’s way too little to make much of, but you’d think if there were anything here, we ought to be able to see it.
Based on only this very limited evidence, Mets fans shouldn’t worry, and Jays fans shouldn’t get overconfident. We know the playoffs are pretty random, and this doesn’t look like a way to be able to see how they’ll go. I can even turn the tables around: while the Mets have struggled some against better opponents, they’ve throttled the crap out of weaker ones. And while the Blue Jays have crushed better opponents, they haven’t kept that up against weaker ones. It’s a lot like platoon splits and hitters — you don’t want to pay too much attention to the splits themselves, because you can probably get the most information from the overall performance. The Mets have done worse than some teams in one area, and better in the other. That which is encouraging and discouraging cancels out.
I can’t stress enough how rough this study is. The .500 line is somewhat arbitrary, and this says nothing about the timing of playing those .500+ teams and sub-.500 teams. And, record isn’t always the best indicator of performance. On, and on. I know where most of the flaws are in here. I also know this quick study in no way encourages me to dig deeper, because I’m not convinced there’s anything there. It’s true that the Mets have played a soft schedule. It’s true that the Jays have played a harder one. It’s true, presumably, that the Jays are better than the Mets are. But what you ought to care about is overall team ability. Don’t worry about how the performances break down by opponent.
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.


There are two theoretical points on this that are worth considering in support of the conclusion, and they overlap quite a bit.
1. One of the nice things about the baseball season is that it’s a nice, big sample. However, when we start parsing it, we necessarily increase perceived variance. We expect there to be more extreme results in smaller samples. It makes sense that some good teams should look AWESOME against bad teams and less so against good teams, and some should be vice versa. This is the nature of the laws of small and large numbers.
2. We tend to think of baseball seasons in hindsight as static things. The season represents all events that occurred within it. However, when we start trying to predict postseason performance based on regular season results, we’re not quite doing this. Teams change over the course of a year. The Mets are a great example. This was their opening day lineup:
Curtis Granderson – RF
David Wright – 3B
Lucas Duda – 1B
Michael Cuddyer – LF
Daniel Murphy – 2B
Juan Lagares – CF
Travis d’Arnaud – C
Wilmer Flores – SS
Bartolo Colon – RHP
That’s not an amazing lineup, but it’s also not horrible. Incidentally, the Mets also played a ton of lousy teams in April.
Fast forward to July 4 and this was the Mets lineup:
Curtis Granderson – RF
Daniel Murphy – 3B
Lucas Duda – 1B
Wilmer Flores – 2B
Darrell Ceciliani – LF
Ruben Tejada – SS
Kevin Plawecki – C
Matt Harvey – RHP
Juan Lagares – CF
That is a considerably less impressive unit. Incidentally, the Mets had a (relatively) difficult schedule in the middle of the season.
And finally, here’s the lineup from September 8 against the Nationals:
Curtis Granderson – RF
Yoenis Cespedes – CF
Daniel Murphy – 2B
David Wright – 3B
Lucas Duda – 1B
Travis d’Arnaud – C
Michael Conforto – LF
Wilmer Flores – SS
Matt Harvey – RHP
This is arguably the best lineup the Mets have been able to run all season.
So, for the Mets specifically, we can kind of sort to explain the difference by saying that for a lot of the games the Mets have played against bad teams, they’ve actually fielded a much better roster. It’s not that the same exact team has struggled against good teams and overperformed against bad teams. It’s that the team has incidentally been made up of better players during the softer parts of their schedule and weaker players during the harder parts.
This is the problem with using most team-wide statistics. They tell you a HELL of a lot more about the past than the future.
It is interesting that you mention this about the Mets, when in fact this is even more true of the Blue Jays. Both these teams turned their seasons around dramatically with waiver deadline trades, but the biggest benefactor by far was the Jays, who went from the outside periphery of the race (actually below .500) to the best team in the 2nd half, first place in the division and fighting for first in the league.
Every Blue Jays series in August seemed to be against another contender for the wild card, and it did not end well for those teams. These weren’t just teams above .500, but teams with something to fight for.
I think the peculiar splits for both teams is all about how the team was put together while they were playing. Both teams became difficult to beat after a few key trades were made.
I think if you want to have a truer picture, look at how their offense plays against the top three pitchers, and how their pitching fares against top hitting. I expect Niese & Colon for the Mets are not their playoff starters any more than Hutchison is for the Blue Jays.
If you look at their 2nd half record, it looks like the Jays do score fewer runs against really good pitchers, but they don’t get totally shut down… and it looks like they kill a starter when they see him the 2nd time, boding well for playoff series. (Look at Pineda, Nova, & Porcello 2nd time facing the Jays.)
The only big pitchers that I saw the Mets defeating on were Strasburg & Tanaka. Not their fault, but they didn’t face that many great pitchers the past few months, and even the ones they beat, they didn’t really knock them around as much as they outpitched them.
At the same time, however, this is how the Mets were supposed to be. Their offense, despite being at its best in August/September, is still not the focal point of their team. The Mets are built off of pitching and it’s no secret.
The Blue Jays, on the other hand, are built around one of the best lineups this decade. Again, no secret.
It’s very tough to use different statistics and performances to determine how much better a team is because of how different the teams are as a whole. I’m not saying that the Mets would be able to outpitch Kershaw and Greinke, or even Price in the future, but it’s worth noting that despite not beating up on that ‘big pitcher,’ they were able to persevere and win because of their staff.
This was my first thought as well. Those are the teams that seemingly had the biggest roster impact during the trade deadline, so pre-July 31 and post-July 31 are two different versions of their teams. If their schedules were very easy (Mets) or tough (Bluejays) in August and early-September, that would have a big impact on these numbers.
“Against teams under .500, they’ve won 67% of their games. Against teams at least .500, they’ve won just 41% of their games. Against the Phillies, the Mets have gone 14-2; against the Pirates and Cubs, they’re 0-13.”
And yet, it’s vital that we reward division winners.
http://www.fangraphs.com/blogs/on-the-inequity-of-the-2015-nl-wild-card-game
Really enjoyed the article. Everyone always looks to find the silver lining in playoff teams before the postseason begins whether it be win % against over .500 teams or the “hot factor” going into October.
I kind of agree that it’s hard to find anything but noise here, but at least here’s another possible argument why. The “above .500” and “below .500” groups these records are accrued against may be very different, from one to the next: not all “winning” and “losing” opponent teams are equally good.
Teams playing in a more polarized league — one with lots of good teams and lots of really bad teams — will tend to have far more of a split, it seems to me, and far more likelihood of getting lots of wins against “losing” teams. This is the NL this year: some utter tank jobs and some pretty solid teams. Teams playing in a league with more parity — with few really bad teams and many sort-of-okay teams — will tend to have less difference, or a wins more randomly distributed across opponents. This is the AL this year.
I’d imagine that 3rd group featured quite a few ’90s Yankees teams, or maybe I just remember them beating the good teams, the bad teams, and everyone in the playoffs with Torre at the helm.
Well outside your sample, but I always remember 1988, when the Dodgers eliminated the Mets in the playoffs in 7 after the Mets had gone 10-1 against the Dodgers during the regular season.
That series still keeps me awake at night.
I think the Phillies did the same thing to the Dodgers a few years before that too.
The toughest stretch the Mets faced was between June 30th and August 2nd, mostly in July. It was all contenders, with a west coast road trip against the Dodgers and defending world champions, and a road series against the team with the best record in baseball.
Many of their best hitters were injured or in a slump. That stretch included the most devastating loss of the season, the rain game. That is where there season could have gone to die.
They started that stretch 0-3, but finished with a winning record, thanks to great pitching. They were 13-12 in July. That stretch, in reality, is what made their season. When they got past it with a winning record overall I knew they would win 90 games.
They beat good teams then. Not before or since, despite better hitting. And lately they haven’t beat anyone. It’s about that starting pitching. Let’s hope they rest up before the Mets have to face Greinke and Kershaw.
It still amazes me that the Pirates are 49-13 against the NL East and West this year but have a record of 29-38 against the NL Central. They are 20-6 against the Mets, Dodgers, Giants, and Nats. As a Pirates fan, it’s a little frustrating knowing that their road will still go through the Cubs and Cards. At least I can say that if they get through them, I’ll feel very confident about the NLCS.
For what it’s worth, the Mets have had bad luck against the Cubs and the Pirates. Those 13 losses include four one-run losses plus two extra innings losses. There have been some blowouts, too, but they’ve lose all the close games to both teams. Then throw in 4-3 against the Dodgers and 3-4 against the Cardinals and I’m feeling okay if they get to the playoffs.
I think those results had less to do with luck than with the fact the Mets were fielding an inferior team prior to August when most of those losses took place. For 10 of those losses, the Mets didn’t have on the roster Wright, d’Arnaud, Conforto, Cespedes, Uribe, Johnson, Clippard, Reed, and Matz. These nine players are superior to the players they replaced and all are expected to be on the playoff roster. Which is why I don’t worry about in-season W-L records vs. the other playoff teams. The only team I really worry about is the Dodgers, even though the Mets won 4 out of 7 against them this season. Greinke and Kershaw are the best 1-2 punch in the game at the moment. Much like Randy Johnson and Schilling were for the Diamondbacks.
I watched the Atlanta feed of a Jays game recently, and the Braves’ announcers were convinced that Toronto’s success had something to do with their strength against power pitchers. They further asserted that since teams with power pitchers tend to be the ones who make the post-season, Toronto was well-placed to take advantage.
I have no idea whether even their assumptions were true, let alone the quality of their analysis.
Braves announcers? Which ones? If it’s Chip Caray, I’m not sure how much credence you can give what he says. All I know is that Harvey and Syndergaard limited the Blue Jays to 1 run over 13 innings this season at Citi Field.
Nice work – it really punctures the idea that you MUST beat good teams in the regular season to compete in the playoffs. Bill James found years ago that great teams aren’t that special when it comes to beating good teams. What sets them apart is that they DESTROY the also-rans. To cherry-pick two examples, the ’27 Yankees were 21-1 vs. the Browns (!), and the ’75 Reds were 11-1 vs. the Cubs. And yet every year we see articles fretting about some team’s inability to beat potential playoff teams in the regular season. It’s not something to worry about.
Time for your daily dose of, “Even if the Mets were 50 games over, they’d still suck, and here’s why.”
I’m mostly joking, but there really does seem to be a run on these types of articles recently. Believe me, we get it. The Mets aren’t as good as their record might indicate.
Did you read the article? It says the opposite of that
Why group the data? Grouping the data reduces variation while all our statistical techniques are aimed at ‘explaining’ variation. But if you reduce the amount of variation by grouping you’ve reduced the chance of observing a relationship unless the effect size is really large. You could use number of playoff games in which they participated (maybe percentage of possible games would be better?) or playoff winning percentage as the dependent variables and their regular season winning percentage against >=.500 as the independent variable. You might run into some problems given the distribution of the dependent variables when estimating via OLS but that can be fixed (via transformations or use of other non-OLS models). Then you could try different cut points on the independent variable . . . that way you are not at the mercy of an arbitrary decision that >=.500 teams are good teams.