This is what I was referring to in all those back jokes-references to various methodological thinkers when writing my other long-winded discussions on research on WHIPs below 1.00 and exceptional OBP years: 

Karl Popper and AJ Ayers, the British empirical school, which is highly identified with free market neo-classical economics of the Austrian school (Von Hayek, Milton Friedman), argued that only empirical research on data that could be quantified and questions that could be falsified could be considered scientific and significant. This remains the dominant way of thinking about method. 

But it is not unchallenged: Thomas Kuhn, in his book The Structure of Scientific Revolutions gave birth to the term "paradigms" and showed that in scientific history, data only has meaning and only counts as evidence to demonstrate an argument's correctness WITHIN a given paradigm. Since the research questions we ask and can ask come from given world views, evidence has meaning only within such a context. The facts never speak for themselves. Advances in science were made, Kuhn showed, only by those working in new paradigms, free from the methods of thinking that limited the questions that could be asked in the dominant paradigm.


Paul Feyerabend, author of "Against Method" went beyond Kuhn to show that in the history of science, advances in knowledge were not only made by those working outside a given paradigm (as Copernicus went outside the Ptolemaic system to create a Heliocentric view), but in fact that often were made almost by chance discovery, Anything goes, Feyerabend taught, since restricting any kind of thought as "unscientific"might be blocking eventual, even unintended avenues to future discovery.
1/13/2016 1:17 PM
Feyerabend's somewhat anarchistic theory of how knowledge comes about might be too anarchistic for some of us, but a middle road between Kuhn and Feyerabend was developed in lectures at Oxford by Imre Lakatos, who said that the problem with empirical studies only is that they fail to realize that the point of studying things is to develop theory - that is new knowledge, some insight into understanding data that is not to be found in the data themselves, but also in the parts of thinking that empiricism wants to ban: pure thought, the experience and cultural understandings of the researcher (or baseball fan), the willingness to think outside the box, and in particular - and this is where Lakatos is helpful to us in looking at these odd relationships - between WHIP and ERA or between Slugging percentage and OBP - that "anomalies" - data that don't fit are NOT to be treated as either making the study inconclusive or as irrelevant to the study and so thrown out as dealing with questions that can't be shown for certain quantitatively, or even as "statistical outlyers" but rather are EXACTLY what a theory and analysis should be looking at and dealing with. Any theory will have cases that don't fit. They are part of the theory, and without them the real world is being expelled from analysis to make it fit. 

So when economists tell us that all sides benefit from free trade, they are looking at income, but not at jobs, so who gets that income is irrelevant. If one country trades raw materials and stays dependent politically and economically on another who instead sells high tech or military goods, in the meantime becoming a military and political superpower that can dominate the world and dictate or regulate market conditions is irrelevant to the model. That one worker is born in the town the factory is in and is a citizen with a vote and legal rights and a union to represent them, and another has come from a desperately poor country to work for less, has no legal rights in the country, no representation and so will be exploited is irrelevant to the model, we look only at wage and production costs and you hire the one who costs less. 

The realization that to say that only questions that our quantitative model can show to be quantifiably verifiable is entirely circular and self-referential (and so invalid) is both obvious to anyone who thinks about it for a moment and also easily ignored since it comes from sources that don't fit into the dominant paradigm, which has defined the only things that can count as scientific. 

 

So for those enamored of these new methods, it makes sense to judge a player by how much better than a non-existent so-called average replacement player - a total abstraction from reality - VORP or WAR - but not to ask whether Lou Gehrig would be better at first base than Wally Pipp.

1/13/2016 1:29 PM
I wrote a conclusion that was important to me to write and explaining how all this fits together but lost internet service just as I went to post it. 

I will try to rewrite it, but I am tired now and it is always frustrating and disheartening to put your all in something and lose it to the ether it is hard to get back to the recreating it. So please stay tuned knowing that my argument here is not fully tightened up but what i wrote to do that got lost to technical problems. 
1/13/2016 1:55 PM
I hate that.  At least you're more motivated than I.  I have a tendency to never go back and retype when a long post gets lost.
1/13/2016 2:14 PM
Thanks. Me too. But I will try, probably tomorrow morning. We are at dinner time, grilled cheese sandwiches (no, not an Italian version), those might revive me. 
1/13/2016 2:22 PM
Your patience amazes me prof.  I'd of been tempted to destroy the monitor or something equally childishly satisfying.................       lol
1/13/2016 2:55 PM
"Don't give him any ideas" - my wife - she really said to answer you with this laramiebob !

I did actually slam the table with my fists. Why don't I learn to write things on a Word document first and then cut and paste? 

Also why can't I throw a big-league curve ball and date Natalie Dormer? (Don't tell my wife I wrote that). 

Also why does the phone always ring when I am in the bathroom? 

I ask myself these things at times like these. 
1/13/2016 3:45 PM
Maybe you're in the bathroom more than you realize.
1/13/2016 5:23 PM
ooh, burn
1/14/2016 9:09 AM

Okay, let me try  to try to explain why I take this stuff so seriously:

 

First, in almost every walk of life we are presented now with a way of thinking that is based on being able to quantify everything we do and evaluate it – and evaluate us, as employees, students, citizens and voters (think of the sophisticated use of statistics by election campaigns of both parties, but especially by the last two Obama campaigns, identifying things like one’s favorite TV programs or favorite online games, then linking that to correlations with likely views on issues, then to creating policy proposals or at least campaign slogans to interest such voters and then contacting such voters at home one-by-one), and of course as consumers.

The whole overall impact of these statistical methods is to isolate us as individuals, evaluate us and then act accordingly toward us – evaluating our employment status or pay, advertising and selling us stuff (Facebook personal data and postings, and Google Search algorithms are key to sales today), and so on.  It is not that the data being generated is not true or correct, but it tells only one sort of story and not others that are equally true. It tells about how we act as isolated individuals, but not how we act as parts of communities, workplaces, teams, classes, ethnicities, organizations, churches, synagogues, mosques or temples, labor unions or social movements, cultures, ethnicities, nationalities. For a system whose advocates see as their utopia the creation of a world with no nations, borders, belongings, cultural differences, etc. but only isolated individuals in a global marketplace (British Prime Minister Margaret Thatcher once said “There is no society, there are only individuals” – of course she still sent brave British men to fight and die in the Falklands War, and I am pretty sure that they thought they were fighting for something called “Britain”), all competing with one another as though they were forms of entrepreneurs, these statistical approaches are very useful.

In education, we see this approach used as part of the two most odious reforms of the past few decades: No Child Left Behind and Common Core, with their emphasis on standardized tests, whose results can be quantitatively evaluated to the nth degree, and which are equally disliked by left and right, by teachers, students and parents. But which are now being implemented in schools in Italy where I live, and across Europe as part of EU directives. The teachers’ lessons are to be individually evaluated, evaluation is the mantra, the sacred hymn of the administrative bureaucrat of today. Was today’s lesson in school or in college useful to the individual student to enable them to make more money in the marketplace? Um, it was on Shakespeare, you know…Yes, yes, but we must evaluate the lesson’s income-generating ability.

For scholars who publish academic articles, the key thing today is to publish only in peer-review journals and only those that have “impact factor” – an algorithmic value that is determined by how many other peer-reviewed journals have cited the articles published in that journal. Want to know why intellectuals for the past 20 years have mostly been talking to themselves, saying more and more about less and less? Because this is the only way they can keep their jobs. You want funding for your scientific research Prof. Einstein? How many peer-reviewed journal articles have you published in the past two years? Um, you know it takes a long time of meditating and working things out to come up with the Theory of Relativity to explain how the universe works. We are not interested in that Dr. Einstein, just the facts, and publish your empirical findings in a peer-review journal and make sure it has impact factor or no money for your work. And so on.

 

I have respect for the brilliance of the Obama campaigns, but this is a long way from mobilizing voters for change by talking with union members at work, church members at congregation picnics, block parties, or other places where Americans come together to be part of a community of other Americans and see politics as about common interest, not about their particular interest.

Which brings me to my two conclusion points:

 

1)      These statistical methods are not bad in themselves, and often tell us stuff we need to know. But alone, they do not only give us only a one-sided view of for example what might motivate people to vote, or what makes a good teacher or student, but they also assist in the project of turning us into those people that the statistics show we are. That is, by isolating our activity on an individual basis, evaluating us as employees, as individual departments of a company (news is losing money, dumb sitcoms are profitable, let’s dumb-down the news or else cuts its budget and put the money into more sit-coms), as individual schools, teachers or students, as players in baseball, it puts us all into competition with one another, and also rewards our punishes us – with employment, pay, promotion, demotion, status, honors, grades, funding – one how well we fulfill or meet the measures of those specific statistical interests. So a teacher that inspires their students, or a colleague that others appreciate because of their ability to mediate and work out differences and conflicts among others on the work team in the office are not appreciated by these methods because these roles can’t be measured statistically. It may even be that productivity – students learn more, the whole team of colleagues works better, the team wins baseball games – improves because of their presence. But by definition that role cannot be seen or measured by these methods. In and of itself that is not a problem, EXCEPT IF WE DECIDE, ACCEPTING THE CLAIMS OF THE ADVOCATES OF THESE AS THE ONLY ACCEPTABLE SCIENTIFIC MEASURES OF EXCELLENCE, PERFORMANCE OR ABILITY, THAT THESE REALITIES DON’T EXIST BECAUSE IT CAN’T MEASURE THEM  !! This is by definition a circular, and therefore illegitimate and invalid logic. That you create a system to measure A and not B, and then when you have an anomaly – Derek Jeter is called a great team leader by everyone who encounters him in baseball, but since his mere presence on a team can’t be measured, only his fielding range, and hitting, it doesn’t show up so you would just as soon have Albert Belle on the team instead, cause what effect could human relations possibly have on humans?  It is not like anyone ever commits suicide out of lost love, or depression, or that some people engage in heroic sacrifice out of patriotism, or inspiration from a cause they believe in. That never happens.  Right?

 

So because these methods wish the other realities that they don’t measure out of existence, and base rewards and punishments, attention and notice and analysis and concern only on those things they can measure, they literally force us into behaviors that make us conform ourselves to the abstract statistical model that they are using to measure our performance.  So that is point 1.

 

2)      The other issue is why these particular ways of measuring baseball performance should have become so important today, instead of decades ago. After all, baseball made do with batting average, home runs, RBI, runs, Wins, Losses, ERA for more than 100 years, and while we have discovered that this or that player was over- or under-appreciated for their time, I think most of us could agree that in seeing Cobb, Ruth, Mathewson, Johnson, Mays, Aaron, Williams, etc. as among the greatest ever, that baseball and fans mostly did pretty well, even if we could quibble with or adjust our own rating of this or that player here or there compared with another. But the idea that OPS or WAR settles the issue of who the best players were in order forever is absurd.  What if Ty Cobb was so stressed out by having to play against black players that he had a nervous breakdown on the field in his rookie season? What if World War Two had not happened and the Cold War never occurred and Ted Williams had played those seasons? What if instead Babe Ruth had been drafted and gassed in World War One? What if there had been a cure for Addie Joss’ meningitis? What if Curt Flood had played today? Or Satchel Paige, or Josh Gibson? Who the f..knows?  Taking players out of their actual team, historical, cultural time and place as a context is silly, and of course a lot of fun. But to think that it can REALLY be done, by WAR, instead of done as part of baseball fans’ perennial way of arguing about the game and enjoying it, as we do at WIS, is counter-productive, which is not to say to ban the use of WAR, but instead  to think that those who like to use it to evaluate players have their method to do so, but the rest of us are not under an obligation to genuflect at the altar of that stat or any other.  There are no clutch hitters? All sorts of new statistical approaches show that. I will tell you what, you get the best hitter you can find using WAR and give me David Ortiz, each of our teams’ will bat down 2 runs with people on first and second and two outs.

 

So why are these statistical methods happening now? Marx – I don’t bring him up for political reasons here but because I have learned so much reading him over the years – would tell us that whenever we see a new set of ideas arising in society it is because the way people work, the social relations of production as Marx called them, have changed. Scientists today are working on the idea that our whole universe is actually a holographic projection (put the keywords “holographic universe” in Google or YouTube and you will see that this is increasingly replacing Stephen Hawkings’ previous view of black holes swallowing all light and information as the main theory of how things work) – then, Marx would say, probably a lot of people are doing kinds of work with simulations and with writing software and the like, maybe making video games, and other forms of work involving the digital creation of simulated worlds. So the thinking about the universe of such a society will be that it is a big simulation, just as in the Feudal Middle Ages people saw Heaven as a fixed hierarchy of statuses – God, Angels, etc. or in Newton’s time they saw the universe as a mechanical device that operated with regularity.

 

So what has changed in the social relations of production in baseball? Free agency and enormous pay for players. Take those two things and a lot of what has happened in the world baseball follows: the incentives for steroids – since to make an extra $5000 in the time of Ball Four by Jim Bouton maybe you would cut some corners, but to make $5 million you are willing to do a lot more. Or the concern to rest pitchers, since they are now a big investment, not just an employee, they become a form of fixed capital expense. No risking them. So managers who believe in complete games, or long relievers or whatever still won’t risk it, because they are breaking company property if they screw up, and property worth millions. And then there is Fantasy and Rotisserie Baseball. John Thorne, the baseball writer has said, correctly, even exactly: “When you root not for a locale, but for a range of individual players on different teams, you are not rooting for a team, you are rooting for an investment portfolio” (The Tenth Inning, Ken Burns’ Baseball documentary, Special Features – interviews on Rotisserie Baseball).

 

An investment portfolio, that is what teams too often see themselves as now, and too often our industrial and commercial companies see themselves as such as well – look up the controversy engulfing Yahoo in recent months and the hedge fund insisting it divest of its core business , or the great report by Reuters journalists on how corporate buybacks have now exceeded money spent by US companies on research and development, on product, or on expanding the business – “The Cannibalized Company” – Google that title and then worry about the future of US Business.

 

Baseball players, like increasingly all of us, are being told to “do what they do just to be nothing more than something they invest in” (Bob Dylan “It’s Alright Ma” – 1965). Free agency means they don’t stay on the same team for most of their careers anymore, so their role on a specific team seems irrelevant to evaluating them as players. They are individual entrepreneurs in a marketplace, or seem to be, so the statistical measures that historically and traditionally measure their role as a team player – wins, RBI, no longer interest, especially since their pay is not based on helping to win the World Series for their team (yes, I am talking to you Johnny Cueto !) but presumably on how well they did according to factors that their agent can show their prospective new team’s GM as an individual player isolated completely (if we buy what the stats are showing)  from their context of playing on a team in the first place. Indeed, just like national borders and laws and regulations in the economy or the global marketplace (the Economist magazine continues to support completely open borders), the fact that players play on teams, just as workers or businesses are based in countries, is seen as an unfortunate distortion of the market, not as the whole point of the economy or baseball respectively.

 

Wait, what do I mean? The whole point of the economy is to make more money, and the national interest is of no relevance and any national government activity is always called “intervention” into the supposedly external marketplace no?

 

Answer: quick, someone tell me the name of Adam Smith’s famous book that founded modern capitalist economics. 

 

So there we are, we have these stats, which are useful in many ways, because players are less attached to teams and make a hell of a lot of money. The need to evaluate players separately from teams is important financially to teams and players. And one could make the argument that it is crucial to winning, except that I don’t believe it and here is my closing argument:

 

We know that so many players are being trained from a very young age in Latin American countries by Major League team baseball academies and then signed to contracts if they pan out because it costs less to train 100 kids in the Dominican Republic than to train 2 in Harlem, Detroit, Houston or Terre Haute, Indiana. This is a big reason why the numbers of US-born African Americans has fallen to so few.  But let’s remember that “Moneyball” is called that for a  good reason: these statistics are about investment portfolios, not players. They are not measuring talent, but cost-effectiveness. The players of course are often among the greatest in the game. But there is no reason to think that suddenly there are no African American players that are equally talented. It just costs more to develop them, whereas in the 1950 and 1960s the former Negro League players were the cost-effective ones.

 

In agriculture, there are at least two different measures of productivity: one is output per worker-hour cost – how much grain is produced in an hour of work and how much does it cost. By that standard, not surprisingly, US large-scale, highly automated agribusiness (automated harvesters, petroleum-based fertilizer etc.) is the most productive. Moneygrain we could call it. But another measure is output per acre (or hectare). And the last time I looked Japanese or Scandinavian family farms were the most productive. They get the most out of the soil, in the real world, not the accounting page. Both are legitimate ways to measure productivity, as is the difference between rice and wheat say, since rice is so productive that you can support the populations of China on it, in a country with limited arable land, whereas wheat needs more room and indeed we, with our vast expanses of fertile land,  grow a lot of it on the Plains. Two realities, both true, different measures.

 

So I am not arguing that RBI is a better measure of how good a hitter is than say OPS – it almost certainly is not, though the RBI records are held by people like Gehrig and Hack Wilson, not by Denny Doyle or Bucky Dent, so, since our real standard for judging ANY system of measure is how well it seems to explain the phenomena that we see in the world around us. If a stat tells us that Bucky Dent was a better hitter than Lou Gehrig, we would rightly throw it out. Well, if a stat tells us that Dante Bichette, or Sammy Sosa, or Jose Canseco or Rico Petrocelli were better hitters than David Ortiz was in 2004 when his OPS was .983 – and if you look for batters with higher OPS in seasons you find these and many others – and so you should want one of them up in 2004 against the Yankees in the ALCS, I would support that statistic ---- because I am a Yankees fan ! Don’t take Ortiz, the data show that Danny Tartabull in 1991 was a better hitter than 2004 Ortiz, are you crazy? Go with Tartabull, a sure winner.

RBI are limited as a method of understanding what is happening, but they also measure actual runs driven in in actual games played, not what OPS measures which is, how effective the batter would have been had on some generic team his teammates been on base with a certain regularity - it does not measure that or ask that, I know, but in a sense, if you want to argue that the hits and walks that a hitter got when no one was on base were as or more important than the ones hhee got when they were on base, that is what you are saying, at least if we care about having that player on a team because we want to score runs. So we are measuring the imaginary runs that should have been driven in or scored, not the ones that actually were. Yet how do we know that this same hitter would have gotten those hits in a different situation - maybe the pitcher was not bearing down because no one was on base in a blowout because the supposedly better hitter - let's call him Mr. OPS  -played for a last-place team, and so got some easier HRs to make his team lose only by 10-1, while Mr. RBI on a pennant-winner had to deal with the best pitchers in the 9th inning in close games with the pennant on the line. And got the job done, but got fewer hits and overall bases than Mr. OPS. What are we measuring really? 

And if we see that Steve Carlton wins 27 games for a last-place team, how hard is it evaluate that compared to say Catfish Hunter winning (I forget so bear with me) 21 for a World Series team? Carlton was clearly better (I cherish Catfish and his memory and Carlton is a weirdo, but facts are facts). Does WHIP tell us that? If it does, great, let's find out how much better, if it doesn't, let's include it as one side of the issue, but admit that seeing Carlton pitch for Oakland in 1972 would have been fun. Heh, let's start an early 70s prog and find out. 

 

I am arguing that theories and methods of measurement are both descriptive and prescriptive – they describe a certain reality, though every method of measurement will leave some things out – Imre Lakatos’ anomalies, and should acknowledge these as legitimate parts of the reality being studied, needing other methods, but present as the limits and boundaries of the knowledge we can get from any given one, and theories and methods of measuring or analyzing things are prescriptive, they are acting, whether intended to or not, to make reality appear as A instead of B, of prescribing a certain set of behaviors that it does highlight as being those that are important for making important decisions.

 

Thinking should not be locked into one or another box. As Gareth Morgan, in the best book ever written on managing companies, “Images of Organization” explains, the metaphors we use for setting up an organization, running one, acting in the real world, have real world effects. Neoclassical economics and the statistical methods that have accompanied its hegemony in the past 35 years certainly has. But if we try to make the world fit into the boundaries of a system of thought, instead of acknowledging that all of our ways of understanding the world are limited since they look closely at some things and not others (think of Google Earth and how you can see cars parked in front of your house or see the whole of North America, but not both – yes you can split the screen but in any case you can only focus on one or the other at a given moment, yet both are “true” or accurate pictures; and b) need to be judged by us, as tools to be used by us as human actors and thinkers, not be our judges or masters by which we are evaluated (“The Sabbath was made for man” said Jesus, “not man for the Sabbath” – human tools and institutions are to serve humans, not the other way around), and experience of the real world is what we want explained. The best method, best statistic, best theory or idea is that which best explains what we see happen in the world around us, knowing that it is not telling us everything and can’t be the only one we use to explain the world.

 

Instead if we rely only on methods that tell us that what we see around us is not happening, or that deny that they can even be judged based on whether they explain what is really happening or not. Think of the endless promises of politicians that globalization and free trade create jobs – explain why selling the same product to someone in Japan should create more jobs than selling it in Kansas City, all other things being equal, and at the same time, if our selling a product there creates jobs, what does their selling products to us that compete with one we make do in Japan and in Kansas City? Yet the idea that trade can also eliminate jobs is negated time after time, so what we all know, that industries have closed down, etc. instead of being explained, is ignored as an anomaly by a method designed to show how much better trade is, not whether it is, and by what criteria, nor are what kind of jobs, how well they pay, what job security and quality they produce, of any measurable interest or significance, and so are ignored as unscientific questions.

 

So the stakes are high, despite the seemingly trivial nature of arguing about baseball statistics, hence my taking the time and effort to go through these things in my own way in these Forum pages the past few days.

 

I for one consider the basic units of analysis of baseball to be baseball teams, seasons, and games, and once we have these contexts, we have the additional unit of individual players to analyze which is fun, and also of great economic importance to teams and players and agents, but also should remain fun. 

1/14/2016 10:47 AM (edited)
Statistics are only ever one aspect of what you can look at when determining better players or teams, etc. However, it is precisely because we can't look at everything at once that statistics are useful and necessary. I can't remember every game for a team in a season and I doubt anyone else can, there are simply too many in baseball. So how do we determine who's better? We can talk about human concepts all we want, but those are ultimately qualitative and thus not truly comparable. As such, there will always be room for discussion. Of course, computers can't handle such concepts at all and thus can only use the statistics -- this limitation should be kept in mind whenever we do simulations here at WIS.

Baseball is the most team-dependent and the most individual-dependent of sports at the same time. The batter that is up is up, the rest of the team doesn't matter -- on this basis, it is the most individual-dependent. And yet, when the batter is not up, he is (unless on the bases, in which case he can make a *small* contribution) wholly dependent on his teammates for success -- on this basis, it is the most team-dependent of sports. Thus, it is completely consistent to look at players in isolation as well as to look at teams as a unit., because baseball does this.

Other sports are team-dependent in ways baseball is not. While baseball players do have to work on certain combination plays, such as the double play, cutoff man, and so on, this is much limited compared to the constant working on run and pass plays necessary in football, or the passing and shooting drills and play designs of basketball, soccer, and hockey.

It's interesting how we discuss what clutch hitting might be. We talk about the most unreal situation possible, and then claim that it is an accurate depiction of the real world in a way statistics are not. "Who would you want to have batting in a critical situation?" Well, you don't really get to ask this question very often in a practical position. I stand by my statement that you want the best hitters hitting in such a role. Now, if the hitters are basically equivalent, at that point there might be one that is better under pressure, but since we're unlikely to be talking benchwarmers, they're both already in the game so you cannot simply choose to have them hitting. Baseball is team-dependent, the hitter is whoever's turn it is at bat. I generally don't believe in pinch hitting for position players unless I have a single position that is very weak compared to all other hitters and I either had a very large bench or was using a DH. As such, if I'm going to pinch hit, the best hitter available is likely the call in a pressure situation -- but unless the pitcher or a very weak hitter is up, I'd rather just let the player that's about to hit, hit.

In fact, it might be more practical to discuss this in a negative way: "Who would you rather not pitch against in this situation" because you might have the option of walking the batter. However, unless Babe Ruth is batting, (and someone much worse than Gehrig is on deck), I'm not likely to be interested in giving a free base, so I'll just pitch.
1/14/2016 12:07 PM
Thanks uncleal, I will accept the criticism of my clutch hitting example as going to extremes.  I will think about how to better illustrate my point. 

In the meantime, here is a perfect example of what I mean by an anomaly:

The 1996 New York Yankees, one of my all-time favorite teams as you can imagine. 

They were 8th in batting in the AL, 7th in team OPS, 9th in runs scored.. They were 4th in the AL in team WHIP behind Cleveland, but also behind fourth place finisher Toronto and last place Kansas City which had the third best team WHIP in the league that year. KC was last in team OPS that year, but they outscored their opponents 786 to 733 and still came in last place - so there is another anomaly.

The Yankees won the World Series that year. Here is Joe Torre's explanation: "I'm of the belief that the game belongs to the players, and you have to facilitate that the best you can. I want them to use their natural ability. If they're doing something wrong, you tell them, but I'd like it to be instructive, rather than robotic.  THE ONLY THING I WANT THEM ALL TO THINK ABOUT IS WHAT OUR GOAL IS AND WHAT THE AT-BATS ARE SUPPOSED TO REPRESENT. AND THAT SIMPLY IS THIS: 'WHAT CAN I DO TO HELP US WIN A GAME?'" - Tom Verducci, "Joe Torre: The Yankee Years" page 15. 

So one question that comes from hearing from Joe Torre is this: if leadership, for example, is not a factor, since it can be shown statistically to exist, why are there managers in baseball? Is there any effect by managers on baseball results? If so, show why and how statistically. If you can't, we must assume that managers do not exist, just as in neoclassical economics we assume that borders (one big global labor market), government and law (regulation of business, creation of currency and its management, building of infrastructure, national defense, protection of property, schools to train a workforce in the first place, etc. etc.) do not exist. 

Or, we can recognize that maybe managers exist because experience has shown that they make a difference. Maybe the difference between being 9th in OPS, 4th in WHIP and winning the World Series. 

CORRECTION: sorry I had misreported one stat - the 1996 Yankees were second in the league in batting at .288, tied with Texas, second to Cleveland. But they were 9th in runs scored, below the league average, and yet hit .293 with runners in scoring position. This is what I mean by anomalies. Paradoxes if you like. I would say that they had clutch hitting, which COULD make a difference because of another factor that made such hitting with runners in scoring position - they were 25-16 in one run games - namely that with Rivera and Wetteland pitching the last three innings, a key hit could make the difference in the game, where it would not given their overall run production under other conditions. Sill, overall it is not a very impressive record, it is in GAMES, which is how the game is played that these performances could make a difference.
1/15/2016 7:32 AM (edited)
Posted by dahsdebater on 1/13/2016 1:16:00 PM (view original):
Italyprof, I think your analysis of strand rates has a glaring hole - it's functionally ignoring luck.  You actually pointed out yourself that most of the "outlier" pitchers with strand rates that don't seem commensurate with their WHIP have relatively low IP totals.  That means small sample sizes, so random occurrences are bound to be more likely.  The question is, how many names repeat?  Do the guys with 100+ IP who make the list actually have unusually good strand rates in other seasons?  If it's really a skill, rather than luck, it should repeat season after season.  I suspect that in most of the cases it doesn't.  To an extent, stranding runners is a skill.  Some pitchers are better than others at pitching out of the stretch.  Prior to the emergence of the big bullpen in the past few decades, the better starters would generally not throw at 100% with a few-run lead until runners got on base.  A few - Verlander and Halladay come immediately to mind - still have the "cruise control" capability well in hand.  But even with those factors, I don't think you tend to see guys have massively anomalous strand rates over a career.  Some are better or worse than expected, but a strand rate of .8+ from a guy with a 1.3 WHIP is almost certainly getting lucky.  I suspect this will manifest as a lack of surrounding seasons with comparable numbers.

I would also point out that it would probably be more informative to look for guys with strand rates anomalously out of synch with their OAV than their WHIP.  Doug Drabek and Nolan Ryan pitched at a similar time and both had a 1.25 career WHIP, but you would certainly expect Nolan Ryan to have a better strand rate.  You have to load the bases before you can walk a guy in.  Bases loaded is a small enough proportion of all PAs against that you can basically assume that most of the time walking a guy is not going to score the baserunner.  And even if you have a walk 4, K 3 inning, you'll wind up with a strand rate of 75%.  If you manage to K 3 the next inning while only walking 3, you wind up with a cumulative strand rate of 87.5%.  So it's hard to walk your way into a bad one.
dahsdebater, a reasonable point, I admit. I want to look at both career numbers and to see how many repeats there are. 

Thanks for the Ryan and Drabek example, it is  a good one. 

I would say two things about the question though: 

First, within the logic of Sabermetrics, we would look to see if we find that something is just luck, because we are, within THAT paradigm, trying to evaluate an individual pitching performer - more than the performance itself note - in accord with the approach involving use of evaluation as a means of isolating individual performances discussed above - and so it is perfectly legitimate to ask if we find repeats or pitchers with consistently high strand rates independent of OAV (good suggestion and point on your part) and/or WHIP. 

Second, however, we are not limited to that paradigm as the only legitimate way of seeing things. IF we find that 10 pitchers a year are "lucky" enough to strand runners to such a degree that it wins a lot of games for their team, that is a real fact in the real world. It happened just as a run batted in happened. To ask "would that same pitcher have stranded the same number of runners in a different year?" is a question that has meaning only under the paradigm mentioned above. Instead, to say that the basic unit of baseball, a game, is often determined - we can for example add up all the stranded runners in games and compare them with the run differential between winning and losing teams in games and see how much effect stranding has - by pitchers who are lucky, but that the numbers of lucky pitchers each year is consistent enough that it may mean the difference between winning and losing. 

If we want to explain the outcome of games, the destiny of teams, how pennants are won, this is useful - the anomaly becomes fuel for re-thinking how things happen. 

Instead if we want to know how much to pay a free agent pitcher, we want to know if he was just lucky. I don't pay any pitchers, I watch teams play games and want to know how they win. And anomalies like a hitter having that .360 20 homer season when career-wise he hits .279 with 11 homers, and the same team having a pitcher who for that year manages to get out of situations with runners on base to a degree way above the league average may determine who wins the pennant that year, who goes down as a legend (how many times out of 100 would Carlton Fisk have hit a HR at that at-bat in game 6 of the 1975 World Series? 

Answer: in 1975, once every 29.4 HR (295 PA, 10 HRs that season), career: about 4 times out of 1000 - his HR/PA rate was 0.38 career. 

But what does that stat tell us? In the real world he hit that home run. It counts. Except as a probability. So if there are 100 quantum universes, in 4 of them, the Red Sox won that game on a Fisk home run. Now we need that guy who was talking about the 1969 Mets in "Men in Black 3". 

1/15/2016 6:08 AM
if you're not a novelist you should make your point in

i will give you sentences
1/15/2016 11:37 AM
I think it's novel.
1/15/2016 11:47 AM
◂ Prev 123 Next ▸

Search Criteria

Terms of Use Customer Support Privacy Statement

© 1999-2026 WhatIfSports.com, Inc. All rights reserved. WhatIfSports is a trademark of WhatIfSports.com, Inc. SimLeague, SimMatchup and iSimNow are trademarks or registered trademarks of Electronic Arts, Inc. Used under license. The names of actual companies and products mentioned herein may be the trademarks of their respective owners.