Wednesday, December 09, 2009

How are the 2009 Toronto Bluejays a good analogy for understanding what climate change is?

On November 11, 2009, I posted this entry, but without any graphics showing the baseball trends. I have republished it here with the baseball trend graphs (and removed the November 11th version).

Although the Bluejays won 9 of their final 13 games, this fact doesn't mean that they had any shot at being in the playoffs. Why? Because if you look at their season record, they won only 46.3% of their games, whereas the Yankees won 63.6% of their games (and went on to win the World Series). In other words, the long-term trend of the Bluejays in the 2009 season was of not being in the playoffs, let alone the World Series. Why does all this matter? Well, it's not because I'm a Bluejays fan, but because it's a good analogue of climate trends:

Even through annual temperatures may appear to have been stable or dropping in recent years, looking only at that short-term trend tells one as much about the direction of climate change (either of increasing temperature or decreasing temperature) as looking at the 69.2% winning trend of the Bluejays right before the end of the regular season: bupkis.

The winning trends of teams over 162 games of the 2009 regular MLB season is what determined which teams would go to the play-offs and eventually to the World Series. Therefore, if we look at a representative segment of the Blue Jays' and the Yankees' seasons, you will note that the Blue Jays had a strong downward (i.e., losing) trend over the last season (R^2=0.8236), while the Yankees had a strong upward (i.e., winning) trend over the last season (R^2=0.7743).

Note: For comparison purposes, I have presented the data of all games from 5/6/2009 (one month after the start of the 2009 regular season) to 10/4/2009 (the last game of the 2009 regular season). The percent-wins were calculated based on how many games each team had won since 4/6/2009 (the start of the 2009 regular season). However, the percent-win values for the first month were not included in the trendline calculation due to the strong effects that the starting percent-win values would have on the trendline (i.e., no team can start with a percent-win value of anything other than 0% or 100%, thus skewing the trendline calculation.) On 5/6/2009, the Blue Jays were playing their 30th game, and the Yankees were playing their 27th game.

In an analogous way, it is the long-term trends of temperature increases and decreases that determine how and how quickly climate is changing. If we look at the past ~150 years, we see the following yearly global temperature anomalies:
The trend seems to be in the upward direction: climate change is going in a positive anomalous direction. However, if you look at just the last few years of temperature data, you see something different:

Here, it looks like the overall trend is that of a temperature decline. However, like baseball, looking only at the last few cases (or games) does not tell you about that team's chances at making the playoffs (the Yankees 8 of their last 13, as opposed to the Bluejays winning 9 of their last 13). In the graphic to the left I have selectively chosen the data-frame, and I have not only selectively focused on only the end of the season for each team, but I have also not included the final three games played by the Blue Jays -- all of which they lost. In so doing, I highlight a short-term trend in which the not-going-to-the-playoffs Blue Jays appear to be much better than the definitely-going-to-the-playoffs Yankees. I even re-calculated the percentages (using the same ingenuous method for both teams) to make it look as if the Blue Jays would be a sure-thing for getting into the playoffs. I even get a much stronger trend than in the comparison above, with a whopping R^2=0.9964 for the Blue Jays (and a winning percentage of 90%!) and an equally impressively predictive R^2=0.9694 for the Yankees (who have a "mere" winning percentage of 64% over this same period). If these games were somehow the critical set of eight games that would determine the entirety of the season, then one would be justified in saying that these trends are significant predictors of the overall outcome. However, this small snapshot of a trend at the end of the 2009 season tells as much about the likelihood of either team reaching the World Series as looking at recent years' temperature data tells one about long-term climate change.

Therefore when a person who stayed in on February 4, 2009 in Ann Arbor, MI said, "See? It's -9 degrees outside! Global warming can't be happening!" That person is as incorrect about drawing a conclusion as the person who went to see the Bluejays beat the Red Sox on September 30 and says, "Did you see that win? We're going to go all the way this year!"

(Okay, I'm rehashing the analogy that Keith Olbermann read on one of his shows in early October - a story from - but it was a good analogy, so I am repeating it here... but with no mention of the horrible Op-Ed by George Will, and no mention of the complete blindness with which he answered the assertion laid out above. However, you can go read the ThinkProgress story and have a chuckle if you'd like.)

No comments: