ARTICLE: Are Baseball Teams Streaky Over the Course of a Season?

[Brian the Automator]

Thanks goes out to our very own rluznski for putting the time and effort in to create this article:

 

Are Baseball Teams Streaky Over the Course of a Season?

[logan82]

I have always said to people that i would rather play a team on a 5 game winning streak than a team on a 5 game losing streak.

[pitchleague]

Looking forward to reading this tomorrow during my lunch hour!

[Oldcity]

Sweet article; very interesting. I'll comment just to keep it at the top a little longer.

[rluzinski]

Just to make it clear, I'm not trying to prove what has caused the Brewer's streaky season. It could just be luck or it could be many other factors. I don't claim to know. I'm just trying to show how streaky even the most consistent team imaginable is. Even though I was looking for it, I was even surprised how streaky some of my pretend Brewer seasons were. As a baseball fan, it's just tough to get a good feel for what amount of streakiness is normal and what amount is over the top. That goes for batters, pitchers and whole teams.

 

Here's the spreadsheet I used for the article, if anyone wants to poke around:

 

Google Spreadsheet

[alorrigan]

While I agree with what you wrote, would it be more accurate to look at every season in baseball history and compare? I know that no one here has that kind of time. Just sort of wondering.

[TerraceReserved]

Russ, how did you simulate the the seasons? I've seen this done before and never known how it's done.

[rluzinski]

I used the Brewers' season as an example of a very streaky season, nothing more. If I really wanted to test the accuracy of my streakiness model, I'd have to test it against many teams and many seasons. All I was really trying to show was that even a consistent team (one that plays at the exact same talent level every day), is still going to appear to be very streaky. It has to.

 

Let me put it another way, if I asked you to write down 162 zeros or ones in a row with the stipulation that there has to be a total of 81 each when you are done and for them to be totally in random order (simulating a team that has a 50% chance of winning every game), I guarantee you that the numbers written down won't be random. For instance, the person would probably never put 5 zeros or ones in a row, even though there's a 98 chance of it randomly occuring. Something being too random is really less random. My excercise is showing "normal random".

 

Russ, how did you simulate the the seasons? I've seen this done before and never known how it's done.

 

It's just a "monte carlo" simulation. Calculate the probability of winning each game and then use a random number generator to 'simulate" each game. If the Brewers have a 66% chance of winning game, 66% of the time they'll be credited with a win. Then, just average the results of running the simulation a whole bunch of times (I used 10,000).

[Nearsighted Views]

When you talk about how much luck factors in to a team's success in a season, I look at a team like the CWS who are essentially unchanged from last year. Their collapse is almost unexplainable. A slight falloff might be expected but I think your model kind of gives some evidence that, to win a division or a WS, you must have a tremendous amount of luck to go along with your talent.

[endaround]

The White Sox collapse was entirely explainable. There's a reason why BP had them winning 72 games.

[TooLiveBrew]

care to elaborate, end? Just referring to players regressing toward career norms, or what?

[Ennder]

care to elaborate, end? Just referring to players regressing toward career norms, or what?

 

The White Sox won the world series in 2005 with a team that had career years almost across the board. It was no surprise they regressed in 2006 and they actually lost talent between 2006 and 2007 so its no surprise they got even worse.

[rluzinski]

Over the course of a season, a team's record often does not match it's true talent level. Anyone who's watched baseball knows that there's a ton of luck involved with winning and losing. The spread in team records in the league is almost always going to be larger than the spread in true talent. Often, if you want to identify the luckiest and unluckiest teams, just look for the best and worst records. Those are the teams that will most likely regress.