I'm attempting to fnd "clues" that identify a streak in baseball. I thnk it's fair to say that streaks exist, but I'm wondering how one could quantifiably identify WHEN a streak is immenint or HOW LONG a steak will last.
It seems I always "jump on board" just as a steak is ending or has ended.
I'm going to spend this baseball season looking a teams streaks and logging data that may be of value later. For those who have spent time on modeling, can you give me some direction as to reliable information that may reveal an impending streak....winning or losing!
MY THOUGHTS:
It seems obvious that scheduling is a big factor. It should be expected that a superior team playing a set of games against inferior opponents MAY result in a winning streak. How would one compare or evaluate a schedule to identify a possible winning streak. ie....Home/Road, etc.
I have data for pitchers and teams that span different time periods. For instance, I have the usual set of YTD data. Secondly, I have data from the past 14 days and lastly I have data for the past 7 days. If you were to use this data, how much weight would you give each set. In other words, is YTD data worth twice as much as current data?....half as much?...how is data collected from the past 7 days more important than YTD data. If you were to put a percentage on Pitcher data, for instance, what % of importance would you give for each of these:
Year-to-date stats
Past 3 starts
Home/Road
Last start
I may be barking up the wrong tree here, but I've got the time and inclination. It will also hone my weak but improving, excel skills during the evaluation processes.
Good luck with your action....