Baseball Analytics Chapter One

Baseball Analytics Chapter 1

Everyone says he wants to bet on “the better team”, but most folks are willing to accept overall win percentage as the quality of the team. That will be correct at the end of the season, but on a daily basis it is not necessarily so. For instance; try to assign a power rating to any team. You can look up Covers rating or get much the same compilation from MLB, ESPN, CBS, Yahoo or dozens of other outlets. There will be a little juxtaposition based on the criteria they used and how they weighted the numbers, but you can bet your last bottom dollar the Cardinals, with 63% wins to date will be 1 or 2 on every list, and the Phillies at 34.7% wins will be at the bottom of all the same lists. That type of power rating is absolutely useless on a daily basis.

The problem is the power rating assigned is an average or compilation of what the entire pitching staff, bullpen and offense have accomplished under any and all circumstances. It does not account for the facts of whether or not your best-worst starter is taking the hill, or what class and type of pitcher your offense is expected to hit against.  

Example: As of right now, on this date, before any games have started, my power ratings say Chris Sale is the best pitcher in MLB (by a razor thin margin over Kershaw and Greinke, but that is NOT what we are here to discuss). So, best case scenario for the White Sox is Sale, at home, versus right handed opponent starter. The power rating comes up 121.6, pretty much competitive with any team in baseball regardless of the Sox overall 46.7% win percentage.

Now look at worst case scenario. White Sox, road, Danks, versus left handed opponent starter = 89.6

So, what relationship does their overall win percentage or composite “power rating” have to do with today’s game? If you answered “absolutely nothing”, memorize this, print this, start a notebook or file, and go to the head of the class to await Chapter 2.

What my program does, via the input of critical stats and results, and weighting those stats and results is to provide projections, and compute the “true” power rating of any given team on any given day. It is not their prospectus for the balance of the season or even tomorrow, and not necessarily a reflection of what they have already done, just what they can reasonably be expected to do TODAY, the day we are wagering on them.

My program also does at least two things no other system or program does. Pitching and offensive stats are weighted and rated for strength of schedule. Who you performed against, and under what conditions, is as important as what you compiled in raw stats. I believe most traditional and SABRmetric compilations rely on raw data, not weighted, but my program will give you more credit for posting 8 hits and 4 earned runs versus Sale than versus Danks. It does the same thing for starting pitchers. For Sale and Danks there is a great deal of difference between facing a pathetic lefty hitting offense like the Twins or a great lefty offense like the Tigers. Pitchers have to be credited for that, just as offenses have to be credited differently for what they did against Sale or Danks. Versus Sale they may lose the game, but gain offensive performance points, whereas versus Danks they could win the game and lose offensive points.

Everything is a matter of perception and computation. I did not have to hold my nose or have giant gonads to bet the Mets Sunday or the Phillies Monday. The Cardinals have a pathetic offense for lefty and Niese has been very good if not spectacular lately. The only surprise was not the Mets winning but that Cooney lasted as long and did well as he did. On the Phillies, how could anyone in his right mind think that with Moore on the hill the Rays should not only be a road favorite, but had a 66.67% chance of bringing home the bacon? This is MLB professional baseball, the best in the world, not your son’s Little League where some teams win them all and others get crushed every game.

That is analysis of true probability, not perception of who is great and who sucks.  

*woosh*

Would be interested to discuss this with you, you're ratings seem arbitrary and I could probably help you fine this. I made a MLA for NFL last year, resulted in 82%.

Shoot me a message when you get a chance, BoL key!

thank you mr. key

No, written so clearly that everyone should understand, and learn something!

That was very enlightening, however I'm confused on one point you brought up. I thought the Twins were one of the best hitting teams against LHP, yet you said they were pathetic. Am I missing something?

Thank You for sharing your analysis and Plays 

Awesome read Key. Thanks for the time 

Thank You Sir

Good thinking and logic Key,....AND,....WHOEVER SAID PICKING WINNERS IS EASY?

LOVE YOUR WORK

One other thing I should have mentioned regarding power ratings. On any given day a team has as many as 20 different power ratings. You have 5 rotational starters, your offense may face left or right handed pitching, that’s 10, and then you may play at home or on the road, which doubles the possibilities again. Football and basketball people are used to 2 ratings, home and away, but that is not the case in baseball.

i guess i am a minority............because i do not understand anything you have written

In a world full of lying touts, delusional fools, and money hemorrhaging idiots it is a pleasant surprise to find the likes of one Key Element. I'm loading up my 5dimes account in preparation for college football and until that spectacular first Saturday come September I'll be following the advise of our very own Key Element in baseball.

Gold star on your test paper! YTD stats read that way, but remember what I said about grading? The Twins had a heck of a streak up to May 24, but are just 4-8 since then and their offensive era versus lefty is now just 2.8 per 9. Complete reversal, but folks using YTD stats are probably not even aware of it. All stats and projections have a shelf life and if you are using April stats in August you are in deep doo-doo.

 for spotting that anomaly

Meth  head  this is a site to discuss  .. ..Im not sure whos head this is over.??/  Maybe yours? If you read what was posted  it does mention that the ranking numbers are subjective ,,position of teams are basically the same .If you have something to add ,it seems weird you want to do it privately .. 

 

Nice write up Key, just wondering if Pitchers Splits, Teams in hitting slumps, crappy bullpens, or Streaks good and bad come into play when capping your game, I can also add ridiculous lines.... Public Favs or Dogs to the list along with key injuries.    

This is great Key, very well written and explained.

Not sure in what sense you mean the numbers are subjective. The only subjectivity is that I built the program, but all entries and computations are mathematical fact, objective, uninfluenced by any method I could impose for personal feelings, biases or influence. It is what it is. I am not betting against it (learned my lesson) and do not want it to treat any team or player in any different manner than anyone else. The program does not know or care that the local team is the Diamondbacks and that I just want the numbers, for or against, so I can wager accordingly.

All or none of those things may have an influence, but to a much lesser degree as they are all peripheral in nature and you must have the core basics on your side before even considering those. Some highs and lows are quickly accounted for in the program because as I told Hog_Wild, stats have shelf life, newer stats must carry more weight and eventually the influence of any stat must disappear. Hitting slumps are accounted for very quickly, but streaks are a day-to-day thing.

I still have to manually input daily, but what I want is available all over and I have set the program up so I can basically copy and paste, then run the computations almost automatically. There is about two hours of boring input, but after that it is only about an hour to decision time. Play, post, wait for the games.

Thank you very much Key. I appreciate the time that you spent on this write-up. I now have a new factor to add "Pitching and offensive stats are weighted and rated for strength of schedule."

Good stuff.