Friday, June 28, 2013

Historical goal trends over the season: why the goals are about to temporarily dry up



A few weeks ago I espoused that the end of the goal drought was here and we would be seeing plenty more goals in the coming weeks. Well, I was right – in the short run. I want to give an updated and more informed idea of what the historical trend is.

I have gone back through this season and each of the six previous seasons and separated the games out into what would be considered a “normal” week for fantasy, meaning I generally grouped games Wednesday through Tuesday into one week (note: some instances required grouping multiple Wednesday through Tuesdays together to get a reasonable amount of games played). These breakdowns won’t be perfect, but they are decent enough for what I am looking to accomplish. I then looked at how far away that week was to the end of the season to adjust for the seasons that have had different numbers of actual games and gameweeks. From there, I created a 5-week moving average for the average number of goals scored for that week based on how many weeks it was until the end of the season. I chose a 5-week moving average for the sake of eliminating the variability from a single week’s potential for all favorable or unfavorable matchups, injuries, international duty, etc.

This graph shows the gist of what has happened, on average, over the last six seasons (excluding 2013). We are currently at 18 weeks remaining in this season, or exactly where the peak is in the middle of the graph. Historically, the goals tend to dry up over the next 5 weeks or so, to the lowest scoring rate we observe. I expect to see a similar decrease in goals for the next couple weeks as our current point for the last several weeks is way above the norm for having this much time left in the season. However, I don’t expect to reach the lows we see from a couple weeks ago, as that was also extremely irregular based on the other six seasons’ data.


I also did the same procedure looking at the 5-year moving average from the start of the season, instead of games until the end of the season. It produces the same general trends: mediocre start, a dip around the middle of the season, and then a rise as the playoff-battle heats up.


Looking at where we are in the cycle yields different results based on whether we want to look at things from the start of the season or work backwards from the end of the season. This is partially a result of the fact that none of the previous 6 seasons gave me more than 30 fantasy weeks, compared to this season’s 36. So, it makes sense to shift both of our current season lines two or three weeks in the proper direction, which would put us right on the first part of the downward slope for goal production.

Of course, this season could continue to be completely nuts and disregard the general trends that the last six seasons experienced. However, there is a significant enough system of peaks and troughs that I want to at least consider that in my transfers and setting my rosters. If my choices are playing my budget forward or my budget defender, I am going to choose the defender every time over the next couple weeks until we can expect production to increase during the final run-in.

Here's what the graphs look like for each individual year.







No comments:

Post a Comment