Quote:
Originally Posted by Yankee Hotel Foxtrot
Thanks again everyone for the very beneficial advice.
I'm having a great time so far.
I originally started out with a small fictional league just to get a feel for things, basically letting the AI manager make most of the hard decisions and all of the game-day decisions.
Slowly I've felt my understanding improve to an extent that I'm taking over more functions and a more hands on role in things like roster management (promotions, demotions), line-ups & draft and trades.
I'm playing the odd game or two myself as well, with some good results and also some not so good.
I then set up a fictional Historical league starting in 1936 and have done really well over the course of 4 seasons(with albeit a fair bit of help from my AI manager) winning back to back World Series in 38 & 39 with the Dodgers. The success has emboldened me to try assume more control for my next season.
I even watched my first game of baseball (Giants v Phillies) in about a decade yesterday on my day off from work. Not sure I gleaned a great deal from it that I could immediately extrapolate to playing OOTP(and being an Aussie I still would prefer to watch 5 consecutive days of cricket!!  ).
One area that I do feel is holding me back some is my nil to limited understanding of most of the statistics (other than the basic ones) and their correlation & significance to strategic or tactical decisions that I as a manager should be making based on my interpretation of the data.
As a result until I can get a better understanding of how to process the data, I think I am going to have to rely on the AI manager to make a fair amount of the decisions.
Anyway, once agian thanks for time you have all taken to help me along the steep learning curve, it's greatly appreciated. |
I could go into a huge ramble right now, but for now I'll keep it simple.
Statistics are either predictive or descriptive.
Quote:
What are the characteristics of a good predictive statistic? A good predictive statistic should be relatively stable from year to year. It should be useful in models designed to predict and plan for upcoming seasons.
What are the characteristics of a good descriptive statistic? A good descriptive statistic should inform about a player's past performance. A good set of descriptive statistics should provide a good feel for how valuable a player has been in the past.
|
Getting a grasp on which statistics are which is extremely challenging, and a quick google search on "baseball stats" should bring up some sort of argument as to which stats are which.
Even the GM's of real life teams don't agree on which statistics have merit.
To put my opinion simply, stats that are context sensitive are not really worth too much. I don't put much value into RBI's because they tell me what someone did (drove in a runner while batting), but they are very context-sensitive (someone needs to be on base, or you need to drive yourself in by hitting a HR).
The best stats (again, in my opinion) are ones that boil baseball down to it's core. How many times, on average, does a player get on base? How many bases does a player take per at bat? Those two things explain a lot, as far as the offensive side goes.
Two stats tell you the answer to those questions. OPS and wOBA. I prefer wOBA because it weighs each outcome correctly in reference to all possible outcomes.
With OPS, a homerun is worth 4 times as much as a single, but in reality, a homerun is not worth 4 times as much as a single. It's a complicated formula and a much more complicated process to determine how much more a HR is worth than a single, but all that work is done for us by some really smart dudes.
wOBA tells us how much a player contributed offensively. Anything under .300 is pretty bad, anything over .400 is very good.
For pitchers, FIP is another complicated stat but knowing where it comes from gives us a good reason to use it.
Most people look at ERA to gauge a pitcher's quality. While it's not horrible, it is again context sensitive. The defense behind the pitcher probably impacts the number too much, so we don't really know what we're looking at without studying lots of other factors.
FIP, which stands for Fielding Independent Pitching, boils pitching down to what the pitcher can surely control. The pitcher can strike people out, he can walk batters, or he can give up home runs. Anything that's hit into the field is partly his fault but also the fielder's. So, they're removed from the equation to make a more reliable predictive statistic.