Win-loss record isn't best indicator of M's future
If water-cooler chat and sports radio calls are any indicator, Mariners fans are ready to abandon ship. But when the smell of panic is in...
Special to The Seattle Times
If water-cooler chat and sports radio calls are any indicator, Mariners fans are ready to abandon ship. But when the smell of panic is in the air, it's always worth looking at some stats that help interpret just how seriously they should be running around as though their hair is on fire or, alternatively, how much hope there is for a turnaround.
The first quarter of a season's games neither guarantees a pennant nor the cellar. It is, however, a decent indicator of likely performance through the rest of the season.
There are two ways to look at the M's record so far. We can check to see what other teams in recent history with similar records through this many games ultimately did. And we can check something useful — Bill James' baseball makeover of the Pythagorean theorem — to see if the M's have been playing any better or worse than their win-loss record is showing.
Through Monday night, the Mariners' record was 15-23, a .395 win percentage. That projects out to a 64-98 finish.
Since 1990, 22 teams have finished their first 38 games with a record of 15-23, and another 23 with records of 16-22. We actually can pool those two groups together because outcomes for both sets are essentially the same (the 16-win teams as a group performed no differently over the rest of their seasons than the 15-win teams did).
Of this group of 45 teams, I eliminated four that played in the 1995 season, 18 games shorter than normal because of an owner lockout. I eliminated two expansion teams in their first seasons.
How did the 39 remaining teams finish their seasons?
There were 30 teams that improved their play, four that continued within a game of that percentage and five that faded from that already low mark. Improvement is much more common than not.
How many of the 39 made the playoffs? Just two. One was the 2001 Oakland A's, who followed their 16-22 start with a blistering 86-38 conclusion, 102 total wins and a wild-card spot. The other was the 1996 St. Louis Cardinals, who had a 72-52 mark the rest of the season, good enough to win the balanced NL Central.
A third of the 39 teams managed to play .500 or better ball for the rest of the season, and just 10 managed to finish the season with a .500 or better record.
The median average team among the 39 played .468 baseball for the rest of the year and finished at 72-90.
To judge if a team's record is a reasonable reflection of its play to date, researchers use Bill James' Pythagorean Theorem. The formula factors the sum of the runs a team has scored in all its games and the sum of how many it has allowed. The results generally are close to a team's won-loss record.
James' theorem usually indicates the actual quality of a team's play better than the team's current record does. Further, teams usually start winning or losing in line with their actual quality of play, so the Pythagorean result is a decent predictor of future results.
And when the formula projects those wins and losses and they're different from the team's current won-loss percentage, more often than not the team will start to win at a percentage closer to the Pythagorean formula's result than the current actual one.
The Pythagorean win-loss tables for the AL West don't change the order of the standings but suggest a much closer race to come (it puts the Mariners just three games out). They bode poorly for Oakland, however. As poor as its record is, the A's have a ratio of runs scored and allowed that hints there's worse to come.
Studying the numbers suggests the M's will play at about the pace preseason predictions posted, though perhaps not at the pace more passionate followers had hoped for.
Panic? I think it would be better to grab an extinguisher and foam down some flaming heads.
Jeff Angus will write a regular feature during the season on baseball statistics, dissecting how they work, describing what they reveal about the game and how that affects the Mariners. He is a management consultant, author of the "Management by Baseball" Web log and the book of the same name.
|PW-Pythagorean wins; PL-Pythagorean losses; RS-runs scored; RA-runs allowed; W, L, Pct. and GB are actual numbers through Monday's games. Pythagorean win percentage = Runs scored² / (runs scored² + runs allowed²)|