None of us could have guessed that this ship would sink so fast. I’m not one normally prone to overreactions, and my goal usually is to temper the highs and lows that the grueling 162-game MLB season brings. "They’re not as bad as they look right now," I could say. "Regression to the mean," I could predict. But that’s not why I'm here today. Sadly, I’m looking to kill your remaining hopes with facts and data. I don't like who I've become - the Mariners have made me this way.
I was spurred towards this research by a discussion I heard on the radio yesterday. The question posed was one that we M’s fans are used to hearing each year for the last eleven years sometime between April and August. Is the Mariners season over already? Are their chances shot? Is it time to resign ourselves to monitoring Felix’s run at a second Cy Young Award?
The answer I’m here to give you is basically, "Yes."
I understand that we’re only 12% of the way through the season, and yes, technically the M’s could still win 149 games. But obviously, just because something is possible doesn’t necessarily mean that it is likely to happen. We all know intuitively that the chances of the Mariners winning 149 games this year is <.0001 but what about our chances of winning games or this is i wanted to calculate.>
This biggest flaw in my research is that I’ve limited my sample size to include only the Seattle Mariners. I’d love to expand my data to include all seasons for all professional baseball teams since 1900, but I simply do not have the time. So what I have then, is a sample size of 37 Seattle Mariner seasons from 1977-2013.
First, I used excel to create a simple line chart showing both Win% in the first 20 games of each season (blue) and Win% for the entire season (red). Just by using the eye test, you can see that these two statistics are at least somewhat positively correlated. It turns out that they have a positive correlation of .744 (1=perfect correlation, 0=no correlation).
Second, I created a scatter plot with "Wins in the First 20 Games" on the x-axis and "Season Win%" on the y-axis. Again, it is clear that there is a somewhat strong correlation here. The two strongest outliers are the 2001 season, which I’m sure most of you can agree was an outlier by most standards, and the 1980 season, when the M’s started 11-9 and finished 59-103.
But what about the 2014 season, right? That’s what we all actually care about now. Well, it turns out that if you compare 20GmWin% with SeasonWin% for all 37 seasons, the mean difference is -.004. What that means is, on average the Mariners’ SeasonWin% has been .004 less than their 20GmWin%. If that held true this year, the M’s would win 56 games this year (.346 Win%). However, you also have to consider the standard deviation of my sample. Excel tells me that the StdDev for these 37 seasons is .105.
Now that we have values for the mean and the standard deviation, we can calculate the likelihood of the Mariners finishing at any Win%! Great, that’s just what I wanted! This is based on a normal distribution.
Let’s get right down to business.
So I predicted that the 2014 edition of the Mariners would finish with 81 wins, so the first test I run is for the M’s to finish at .500. I get .0071. According to the data, the M’s have a 7.1% chance of finishing with at least 81 wins. That’s not so bad, I guess...
But what about the playoffs? The hope of making the playoffs is what makes the MLB regular season meaningful, so it would follow logically that the Mariners would already be playing meaningless baseball if their chances of making the playoffs are very low, let’s say less than 5%. And even though 90-win teams don’t always make the playoffs, let’s be generous and say that 90 wins would somehow be enough to get us to the playoffs this year. My test spits out a value of .0023. That means the M’s have only a 2.3% chance of reaching 90 wins in 2014.
This is indeed a disturbing universe.
Let’s go Felix!