As my toddler grows, I know the random questions will begin. Why is the sky blue? Why can’t I drive the car? What is the point of soccer? Why is that man bald? I know that I need answers for these, and they need to be simple. For example, soccer is easy to explain. “Sweetie, ignore the crowd chanting and the players flopping around on the ground. The important thing is to get the ball in the goal more than the other team without using their hands.”
Baseball is simple, but with lots of nuance. You must hit, pitch, and field well, and have the depth and strategy to be able to do it again tomorrow in order to be successful. But that’s a long and complicated answer. What if you could boil baseball down to its elements and pick out the one thing that is most important? “Sweetie, ignore the men scratching their pants and hitting good players with baseballs without really getting in trouble. If we hit the ball over that fence more times than the other team, then we will get to play fun games in October.” Now, baseball is more complicated that that, and I am glad that it is. Every player on a good 25-man roster has a specific and important role. It’s why assembling the correct 25 is such a tough task versus just seven or eight (if that) in basketball. But the current run environment and the launch angle revolution is making home runs more and more crucial to a team’s success.
Just how important are home runs to the outcome of a game? What if we could remove walks, base hits, strikeouts, groundballs, flyouts, infield hits, infield flies (yes please), pitching changes, lineup decisions. All of it. Hitting more home runs is a win, hitting less is a loss, and ties don’t matter. Would it alter the course of our playoff race? Here is the National League East standings: (All numbers were crunched through August 22nd)
Home Run Standings | ||||||
---|---|---|---|---|---|---|
Team | W | L | T | Win PCT | Prorated Wins | Prorated Losses |
Atlanta | 52 | 39 | 35 | .571 | 72 | 54 |
Philadelphia | 52 | 42 | 33 | .559 | 71 | 56 |
Washington | 48 | 45 | 34 | .516 | 66 | 61 |
New York | 37 | 47 | 42 | .440 | 55 | 71 |
Miami | 33 | 58 | 37 | .363 | 46 | 82 |
This looks a lot like the real-life standings. What’s more is that the home run-based W-L-T approach gets much closer to approximating the actual win percentage than the Pythagorean method of using runs and runs against to determine the winning percentages of the division.
For the uninitiated, the Pythagorean method uses the total runs scored and given up over the course of a season to determine a winning percentage. It’s quite accurate and is used to filter out luck relative to the actual strength of a team. A team’s Pythagorean record is not always going to reflect its record at the end of the year because the season has a finite end, but if the season went on infinitely it’s probably a safe bet that the team’s actual record would approach its Pythagorean record. You can find how to calculate it here and I will be using the Baseball-Reference method.
So I compared Major League’s HR standings, Pythagorean method using runs and runs against, and also the Pythagorean method using home runs and home runs against. Did it beat runs scored Pythag runs MLB-wide? Well no, but looks how close it came.
Side-by-side comparison by how many games away from actual standings | ||||||
---|---|---|---|---|---|---|
Method | Average | Median | StDev | Mid 87% Avg | Mid 87% Median | Mid 87% StDev |
HR Standings | 6.2 | 5.5 | 4.50 | 5.8 | 5.5 | 3.37 |
Pythag with HRs | 6.1 | 4 | 5.27 | 5.7 | 4 | 4.18 |
Pythag with Runs | 3.4 | 3 | 3.00 | 3.1 | 3 | 2.21 |
So by just filtering out everything but home runs, you can get within .037 of the actual win percentage (6.2 games) versus Pythagorean with runs getting you within .021 (3.4 games). This kind of makes sense that it would be about twice as accurate, since homers account for about half of all runs. Filtering out the top two and the bottom two teams (rightmost columns) helps a little, getting you within .035 versus .019. Let’s look at some of those outliers.
White Sox - HR standings Win Pct .516, actual .381
Orioles - HR standings .426, actual .291
Tigers - HR standings .312, actual .409
Mariners - HR standings .476, actual .563
The teams that this model fits the worst are the ones playing the worst baseball... and the Mariners. So there might need to be a scaling adjustment made for above .580 and below .420 teams, because it couldn’t keep up with the Red Sox very well either. And if you can explain the Mariners defying all baseball logic, please meet me at the track.
But let’s go back to the NL East, with pro-rated HR wins and losses to account for ties, the HR standing win percentage, Pythagorean win percentage, and actual win percentage.
NL East | |||||
---|---|---|---|---|---|
Team | HR Wins | HR Losses | HR Win PCT | Pythag Win PCT | Actual Win PCT |
Atlanta | 72 | 54 | .571 | .573 | .563 |
Philadelphia | 71 | 56 | .559 | .501 | .540 |
Washington | 66 | 61 | .516 | .554 | .501 |
New York | 55 | 71 | .440 | .468 | .444 |
Miami | 46 | 82 | .363 | .367 | .398 |
If Pythagorean win percentages we’re happening in real life, the Nats would be right behind the Braves, the Phillies at around .500, the Mets a little closer to the pack, and Miami bringing up the rear regardless. So why hasn’t this played out this way in real life? HR standings get you within .016 win percentage (2.6 games) of actual outcomes versus .029 win (4.8 games) with Pythagorean. Are home runs more closely related to NL East teams’ success than in other divisions? Maybe so, the Braves seem to get a lot of juice from them.
But it might account for the Phillies outpacing the Nationals. Washington has given up at least two home runs in 42 games and lost 26 of those. They were outscored by .46 runs per game in those 42. Philadelphia has given up two home runs in 34 games and lost 25 of those. They were outscored by 3.32 runs per game in those 34. So when Philly gave up multiple home runs, they were blown out. Washington lost one more game when they gave up multiple home runs over the season, but those games were closer.
Philly has given up zero home runs 54 times (second best in baseball), winning 43 of those while blowing out opponents by 2.7 runs per game (third best in baseball). Washington has given up zero home runs 40 times, only winning 26 times while winning by 1.5 runs per game (16th and 18th in baseball).
I bring up all of that to say this: the Phillies win big and lose bigger than the Nationals (which can be seen in Pythagorean win percentage), but the difference comes from their pitchers keeping opponents in the park. Now, the fact that the Phillies have more home run wins and more games with zero home runs given may be a symptom rather than the cause of their winning. Basically, the Phillies might be winning more games than the Nationals simply because their pitching is better (there is a .30 gap in FIP). They can dominate single games and opposition home runs as a result.
Home runs aren’t everything. Of course you need pitching, preferably pitching that gets strikeouts. It doesn’t hurt to have two great young defenders in the outfield for those types that the strikeouts don’t happen. A good pinch-hitter and a deep bullpen would be nice to have. But look what happens when you remove everything but home runs. Baseball looks pretty similar. It seems to revolve around homers now, and if the Braves can keep hitting them and while keeping opponents’ fly balls where Ronald Acuna Jr. can grab them, it will make the season and the playoffs so much easier.