State of the League

 

Look at that graph.  Look at those 60% transparent baseballs in the background.  Man (and Justine), that is a goddamn fine looking graph.  So colorful.  So easy to compare.  So actually conclusion suggesting.  Isn't it amazing how, after Oksana went to sleep before the Jacuzzi in weeks 6 and 7, the quantitative categories go down and the ERA and WHIP rise?  And then fell in tandem with the rising other categories?
I feel sadness that the league was in such a slump in those weeks 6 and 7.  I feel happiness that moves were made and the league is now as potent as it evar was.  I feel pressure that the league is doing its best lately.  I feel worried about how this week 15, where there will be a full week plus 4 post all star break days, will affect my statistical correlations.  But then I look at that graph again and I feel happy.

Materials and Methods:

Team totals were tabulated and the averages of each category for the entire season was calculated.  To calculate the individual weekly totals, a more complicated moving average was required to smooth out some of the noise in the data.  Thus, for the meat of the weeks, the team averages were calculated by multiplying the team's performance in that given week by 4, the preceding and next weeks by 2, and the weeks surrounding that 3-week period by 1, and then dividing that by 10.  For the first and last weeks, the same calculation was used except using 1-tail rather than 2 in the weekly tabulations.  To construct the graph above, all data needed to be scaled so that it could be accurately compared (40HR can't be compared to 0.300 average in the same units).  Thus, a variance-based approach was used where the weekly floating average was subtracted from the total average, and then divided by the total variation (the difference between the maximum weekly average and the minimum) to give values that range from 0.3704 (AVG in week 14) to -0.3198 (K/9 in week 7). 

Worst Week Evar!

The commish recently noted in his power rankings that Sabo's team was coming off of a bad week, and to now focus on moving forward. This got me thinking about what the effects were of an especially bad week and how badly they hurt your team's overall record. My initial thinking was that it was most important to limit your damage on a bad hitting week by pitching well (and vice versa), and that to make up any sort of good ground in the overall rankings, you would need to make your great weeks count by taking 9, 10, or even 11 categories.
I went through the scores of each team in each week and quantified how good or bad each week was, and how the overall score of that week ended up (See Materials and Methods at the bottom of the post). I will highlight a few specific cases, but check out the original results, find your team, question the data!
1.  The Peachz - Lemonade Makers
The Peachz have been frustratingly resilient throughout the years;  you see their draft and laugh, you see some early pickups and you laugh, you see their record after the first go-round in the league and you laugh, and then somehow, the end up in the playoffs. Perhaps a look at how they performed in their best and worst weeks would give a clue:

Indeed, the Peachz were able to compensate for their bad hitting in week 4 against the Fooeys with dominant pitching to take the week 7-5, and they were able to outslug the Moraleseseseses in their worst pitching performance in week 1 and take advantage of the former Andruses' equally poor pitching to take the week 8-4.  However, they were unable to convert their great power hitting numbers in week 9 against the Omars to more than an 8-4 victory, and did even worse when their usually dominant pitching was unable to outslug the Kurt Locker and were only able to eke out 7 victories.
What really stood out to me when looking over the Peachz records in these matchups was their consistency.  Despite poor performances, the Peachz were able to add wins regardless of whether they were slugging and pitching at their best or were strug-gull-ling.  In other words, the Peachz can win ugly or win pretty, but they consistently win, or at least don't ever drop an insane amount of games (except for this last week, where is Judd coming from?).  Indeed, that is the key to their dominance in the league:  they always manage to get just a few more victories than you do and scrape together wins in various categories each week even if they are unrelated.  If you average 7 wins per contest, you will do quite well in the league.

2.  The Bernabes - Follow the Guitar
If the overriding theme of the league and key to making the playoffs is to consistently win and inch your way into the bracket rather than win big occasionally and then lose the rest of the time, then the Vaya Con Bernabe team is the perfect and tragic illustration of what happens when you consistently lose and you fall farther and farther behind.  The Bernabes are an excellent collection of players with some of the best pitching in the league, but they just don't have that locker room chemistry that carries teams over the top. 

The Bernabes were not able to make up any ground with their ridiculous pitching when their hitting failed in week 8 against the Peachz, and they were unable to smash the ball when their pitching failed (but still won most categories) in week 5 against the Moraleseseseseses.  However, even when they both hit the cover off the ball and pitched the best that they ever had in week 10, they were unable to make up any ground against the Upton Bros, falling to them 5-6-1.  Looking at their weekly totals, it just seems like their great weeks don't match up with eachother:  when they get the most RBIs, they can't score any runs; when they steal the most bases, they can't get any ding-a-lings.  Even their vaunted pitching, so dominant, so many hitless and perfect innings, is unable to match up:  with wins come high ERAs; where Ks are to be found, saves have run for ze hills.  This team is a good team, they are just totally out of sync, which is causing them to drop games all over the place.  Like the Peachz, who consistantly win and climb in those standings, the Bernabe are consistently losing and are thus falling farther and farther off pace.
3.  The Complete Lack of Moraleses - All over the place
If the Peachz are climbing with consistent winning and the Bernabes are falling by losing the same way, the Moraleseseseses are all over the place.  One week, they are losing 1-11, the next they are winning by the same score.  Looking at how they have fared on their best and worst weeks, things become even more murky.

At their worst, they have dropped significant games.  They dropped a total of 21 games against Omar and Ackbar when they hit and pitch poorly, respectively.  Conversely, they took 8 and 8 against the Uptons and the Fooeys when they hit and pitched well.  Currently, they sit second in the standings, and the majority of their games are nice 7 or 8 victories.  They do, however, have these occasional but not rare major stumbles.  Indeed, these types of wildly inconsistent performances have lead the Moraleses to hang with both the Matsuis and Bernabes in the basement, but also to hobnob with the Ackbars and Omar on the rooftop.  How will this team play out?  Will they be able to level out or will they just flame out and leave their crushing victories as little more than footnotes in a wasted season?  We shall find out in about 10 weeks and, in the interim, all pray that we catch the Moraleses when they put their shoes on the wrong foot, rather than putting those shoes up your ass.


For most teams, the worst weeks come with about 4-5 wins, and the best come with 7-9 wins, but the common thread is that most of these performances are not the outliers; there are usually higher win totals when you perform worse, and lower win totals when you performed better.  The key that I found when looking over all of these performances is the best single-week performances are rarely coming from the teams that are doing the best in the rankings; rather, the teams that are at the top are winning most of their matchups by small margins, but winning often.  Just like the real sport of baseball and not this D&D version, teams that win 100 of their games 3-2 will end up going further than teams that win 60 of their games 15-0, but lose the remaining 100.

Materials and Methods - Each team's weekly scores were scored based on how they stood in the ranking of all of that team's scores.  The best performances in a given category were given a +3 score, the second best a +2, and the third best a +1.  The worst performances were given -3, -2, and -1 accordingly.  All other scores were given a 0.  The total hitting and pitching scores were tabulated (column AC "hitting" and AD "pitching" on the right in the "raw data" tab of the spreadsheet), and the best and worst performances for each team were isolated.  Analysis was followed by coffee, and then I concluded on the pot.

The 8-pitch limit

There has been much talk (well, some talk.  Not as much as a David Wright gay post gets.) about the teams that go over the 8 start limit.  I decided to look at this problem from a more statistical point of view and see what the real effects are of going over the start limit.
First, I simply wanted to see who committed this most foul of sins most often. There have been 22 instances where teams have gone over the start limit, and 16 times where one team has gone over while the other has stayed within the limits (3 matchups where both teams went over).



Sabo is weighing in as the most starts over the limit with 5 instances over the limit and 11 total starts over the limit.  Sabo started 11 against me (as well as 11 against 1 other person, and 10 against 2 people), and I actually emailed him earlier in the week and he reveled in his limit-busting roster.  As a lawyer, I think that Sabo goes with the idea that if the law allows it, it is legal, which I guess is sort of true.  Loopholes are legal.  Don't hate the playa, hate the game.  I rest my case on your face, councilor.  Good job Kurt Locker, Ackbar, and Andrus, you are going to fantasy baseball heaven.  
It is all well and good for us to demonize the people who go over the limit, but was it really so bad to go over?  To answer this question, I had to look more specifically at the effect of going over the limit.  I went in with a number of assumptions:
- The more starts, the more Ks, and the people who went over the limit would be winning that category more often than losing it.
- Same with wins.
- Because you have so many more starters, you should not have as many closers and should thus be punting saves.  
- Because you are throwing out more guys, the chances of you starting a crap pitcher or having a poor performance should go way up, so the ERA should be higher when you go over the start limit.
- Same with WHIP.
- Same with K/9.

I looked at how many times a team that has gone over the start limit has taken a category when their opponent has stayed within the limit.  The results were not what I was expecting.



First of all, the way that I read these results is that the extra starts would be said to have an effect on a given category if there was a significant shift away from 50% wins, 50% losses.  If half the time you won and half the time you lost a given category when doing something, it makes sense that that something would really not have an effect on how well you do in that category.
So as you can see, when you went over the limit and your opponent did not, you won or tied Ks and Wins much more often than you lost (75 and 68.5%, respectively), meaning that the extra starts definitely strongly affected these two categories.  However, these extra wins did not seem to come at any cost, as I had initially predicted.  Teams that trotted out all of these extra starters did not sacrifice any of their relief efforts, and the ERA and WHIP remained steady in the face of so many extra innings pitched.
So it seemed like Ks and Wins were the only category that were responsive to these extra starts.  To confirm this, I normalized these categories and saw what the effects were.  I thought it unfair to simply cut out the best performance or the last pitching performance of the week, so I simply found what the average start of the week was for these teams and averaged them over 8 starts.  As expected, Ks and Wins fell into more randomized ratios, where teams won half their games.  


Indeed, had these people followed the rules as they were, they would have given up a bunch of wins.
All of this got me thinking about the effect of extra starts.  Indeed, it seems like if you start more guys, you get better results.  However, if you stay under the 8 start limit, is there anything wrong with that?



Apparently, the winning strategy is just to load up on starting pitching.  The more starters that you trot on out there, the greater the chance of you winning those critical quantitative pitching categories, and seemingly, there is no detriment to the average pitching categories.  Indeed, it is questionable whether having extra bench hitters to fill in on those Mondays and Thursdays will have nearly the benefit of having an extra starter, regardless of quality.  Even if you say that more starters means less quality, you will still be able to rack up those innings and absorb a bad pitching outing more easily than if you had pitched fewer innings.  
One of the interesting conclusions that I came to when looking at this data was that these extra starts are often accumulated at the ends of the weeks, after you already know what your pitching line will look like.  If you had stunk up the joint and your WHIP and ERA is already raising the roof, and not in a good way, or that your opponent's is, then it is very easy to push those starters out there to get Ks and Wins without worrying too much about their actual performance.  I think that we have all been in the odd position when you are happier to have a guy hit a homer than to ground out, simply because you already have sealed up that category.
In the end, I agree with the decision made by the commish and support further the notion that we should just drop all upper pitching limits.  We always make the decision of whether the extra Ks or Wins will be worth the chance of giving up a million more runs, the only crux is that at the end of the week, these averages all set up and the team that is ahead in ERA and WHIP has much more to risk by sacrificing one precious move, a player who they probably wanted to keep, and the sanctity of those low average categories for the marginal benefit of beating your opponent in those two critical quantitative categories.   We may think that it is cheap for a guy who has a 6 ERA and a 2 WHIP on Saturday to pick up a bunch more guys to win Ks and Wins, but it is in fact the same calculated risk we all make except for the fact that a bad outing will not be nearly as detrimental.  Hindsight is 20-20, and most of these decisions are made on the cost-benefit analysis that we all run before the starts.  It seems like you simply have to make the decision that we all knew intrinsically, that you will be getting a few extra Ks and possibly a win, but you are running the risk of a bad outing.

Great Performances - Part 2: Pitching

While individual hitting performances may have not had the transformational effect on the overall weekly score, surely amazing pitching performances would be more important.  After all, we are (in theory) limited to 8 starts and even in this frontier league, it is rare that we get more than 12 starts.  It is reasonable, then, to assume that one of these 8 or so starts being so great would have a heavy influence on the score, right?  Wrong!  For the pitching, I calculated the final score based on the premise that the pitcher had been on the bench that day, but also on the premise that the pitcher had turned in a more pedestrian performance to the tune of 6IP, 3ER, 5K, 4H, 4BB. 

1.  Roy Halladay pitches the 20th perfect game in 123 years of MLB history.

Doc Halladay was dominant against the Peachz over Memorial Day weekend, but his performance only lead to a 1 game swing in the standings.  Perhaps not surprisingly, this swing came from the change in WHIP, but the 9 scoreless innings had no effect on the final standing.  However, even if Doc had scattered 6 hits throughout his 9, the Bernabes still would have had enough of a benefit to take the category.

4.  Josh Johnson strikes out 12 and wishes Vaya Con Dios to Bernabe

Not only did Johnson's 4 baserunners in 9 carry the WHIP category, but his high K total was enough to turn the total K category.  Had Johnson gotten only a single K per inning, which is a nice line to begin with, the category would have resulted in a tie rather than a win for Omar.  In this league that rewards Ks above all else, a dominant K performance is the most important factor in the overall scoring.


5.  Ubaldo Rises to national prominence by no-hitting the braves. 

The no-hitter is, in my mind, the most overrated of fantasy achievements.  Despite giving up 0 hits, Ubaldo gave free passes to 6 batters which, though small in number, was still large enough to add to the W + H total and leave Bernabe a bit behind.  Had Ubaldo given up a single homer in the game rather than walking 6, WHIP would have gone his way that week.  I remember this matchup and, after the no-hitter, Tim was ahead in many pitching categories, but Gavin Floyd's poor pitching the next day erased all good that the no-hitter had done.
The average pitching swing that these great performances had was 0.8 games, less than the 1 full game that a great hitting performance had.  I think that the final line under all of this is that, because there are more easily attainable quantitative hitting categories that are spread over a larger amount of players and opportunities than pitching categories, the great hitting performance ends up mattering more than the great pitching.  Also, bad pitching can be so much worse than bad hitting because of the damage that bad pitching can do; 8 ER in 3 IP is much worse than 0/5.  I also think that it is worth noting that, without fail, the "alt" pitching performance of 6IP, 3ER, 8BR, and 5K is just as good as no pitching performance at all.  Perhaps that will figure into the currently ramping up argument about the 8 starts limit. 

Great Performances - Part 1: Hitting

Perhaps one of the most heartbreaking moments of a fantasy week is when, after being away from your computer for a while, you check how your matchup is going and you see that your opponent has pulled far ahead of you. Looking more closely, you see that one of his (or the Peachz'z) playerz has accounted for most of that damage through a monster day.  Especially when those performances are early in the week, you feel like someone has just punched you in the stomach and, while you are bent over with no breath, made sweet love to your spouse.  I have been in the dugout for my precious Ackbars when I saw Adrian Beltre go yard twice, and when I saw Ubaldo rise to national prominence by pitching his no-hitter in April. At the end of both of those matchups, though, I found myself comfortably winning, leading me to question the value of those single great heart wrenching performances.
Rather conveniently, ESPN has put together their 20 best single game pitching and hitting performances of the year.  I looked at these performances and came up with how much of an effect they had on the final weekly scores in our league.  I should note that, of the 20 pitching and 20 hitting performances, our teams only owned 9 of each when they went berzerk, perhaps highlighting how exceptional these performances are and how anyone can have an amazing day any time regardless of whether he is routinely a good enough player to be noticed by the league.  I will highlight a number of performances, but you can click here to go to the spreadsheet yourself where all 18 of the performances are calculated.  The sheet totals what the team's numbers were with the great performance, without it, what the opponent's numbers were, and what the differential is assuming that the player went 0-fer for that day.  Also, on the right, you will notice the amazing performance and the amount that they contributed to the overall numbers with dOBP, dSlg,  and dOPS being (as scientists and mathmagicians know), the difference or "delta" of those categories:


1.  Adrian Beltre goes yard twice and knocks in 6 against the Ackbars.



As you can see, if Beltre had gone 0/5 instead of 4/5 with all those extra numbers, there would have been a 2.5 game swing in the final weekly total.  It is interesting that all of the categories that Beltre's performance had an impact on were the quantitative categories rather than the average ones.  With the average categories, your other players have a much greater effect than they do on the quantitative categories; the great hitter can carry your team with R, HR, RBI, and SB while the rest of the team slacks off, but that slacking will have a much more drastic effect on the Avg and OPS.  Truly, Beltre was the life preserver for this week for Woo Woo.

4.  Miguel Cabrera rings the bell 3 times against the Orlandos/, /ICU last weekend.



Miggy's amazing performance against the Orlandos was entirely wasted, unfortunately, as even if he had come up hitless, the pre-Kendry-disaster-Andruses would have still walked away with all 6 of the offensive categories that week.  Judd's team had built up a large enough lead in all of the offensive categories that, while dealing a psychological blow, had no impact on the final numbers.  It is interesting to think about how if I had said that Miggy was on the bench that day, Judd would have surely been cursing himself, but in the final reckoning, it would have made no difference.

3.  Albert hit 3 bong-bong-bongs to celebrate the armed services against Woo Woo.



Albert Pujols was coming off of a rough May, but finished strongly with these 3 dings.  While his performance that glorious Sunday was terrific, the overall swing in the score for that week was only 0.5 games, coming from the 3 runs that he scored as a result of his Hamas.  It is interesting that, while I was initially quite excited when I saw on my iPhone MLB at Bat that "A Pujols Homered off of J Grabow," the fact that none of my scoring categories turned over tempered that feeling.  Indeed, Adam Wainwright's win in that same game was the more important stat.  I would have easily traded all of those moon shots for a few more runs on singles by my team throughout the week. 
The average differential with and without these great performances was 1, which is much less than I was expecting.  Of the 9 total games swung with these 9 performances, only 2 were in the "average" type category highlighting how little one performance can affect your averages.  With between 250-300 ABs per week, it is not so surprising that 5 or 6 would not have a huge effect, but I would have thought that the matchups were in general close enough that 3 or 4 runs and 2 or 3 ding a lings would have a big difference on the final numbers.  Hmmmmm, average margin of victory.....that sounds like a good future topic!
Great pitching performance analysis coming, shorty.  I mean shortly.

Tin Can Alley

Reading the commissioner's unnecessarily mean power rankings from this past week (hey, he is my friend, what am I supposed to expect? Friends are mean to each other, right?), I noticed that he said that his team had hit a string of tough week with teams putting up their best numbers against him.  This got me thinking about how a team can really put up great numbers against one team one week, and then stink up the joint the next.  I had previously tabulated the totals of all of a team's opponents and then ranked which team had come up against the greatest cumulative buzzsaw (Omar's opponents had been the toughest in total, followed by the Peachz and FuE, with Andrus' team facing the easiest competition, followed by the Ackbars and Woo Woo).  This week, however, I came up with the list of which team has put up the best numbers in the league in a given week, which team was their opponent, and against which team teams put up their best number.  The results are given below (Click here to go right to the spreadsheet and look at the raw data yourself.  Make sure to check out the tabs). 



The first column, "Best of the Week" shows how many times a given team has put up the best totals for any given week.  As you can see, the Ackbars have put up the highest numbers in a week 19 times, many of those coming in the "Steals" and "Saves" categories.  Conversely, the ICU has only put up the best numbers of the week 5 times.  Given the injury history, this should hardly be surprising as the team has not yet started to fire on all cylinders.  It is interesting to note, though, that despite never killing a given week, this team is still in contention, showing that you don't need to be the best during the regular season, just good enough to beat your opponent.  Most surprising to me was the fact that the Fooeys and the Matsuis have put up the best numbers in a week 9 times each, but yet remain in the cellar in the overall standings.  Perhaps there is a good explanation for this...
The second column of the data, the "Unluckiest of the Week" shows how many times a team has been facing a team that puts up the best numbers for that given week.  For example, there have been 14 times when the Bernabes' opponents have put up the best numbers of the week in a given category, and 12 times when the Fooeys's opponents have dominated the week.  Bernabe's numbers are only slightly ahead of the rest of the field, though it seems that the sight of their unis send their opponents into the white hot rage of a thousand suns.  Conversely, the underperforming Andrus's (AKA the Lack of Moraleses) have gotten pretty lucky in that their opponents have only put up week-dominating numbers 6 times. 
The third column, the "Punching Bag" column shows how many times an opponent's team has put up their best numbers of the season against them.  For example, the Car Wash (or Super Upton Brothers, as they are known these days) has been the subject of opponents putting up their career best numbers in a given category 19 times.  The difference between this listing and the previous one is that this listing measures when an opponent puts up the best numbers of his career, while the previous one only measures when an opponent puts up the best numbers in the week.  Because there are 12 categories and 10 teams, there are 120 chances that a team can put up their best numbers against you, while there have only been 8 weeks with 12 categories, leaving 96 chances to be the subject of a team mauling you by putting up the best numbers of the week.  The bottom line from this analysis is that the Car Wash's opponents have been the strongest against him, and that the Woo Woo's opponents have not been particularly strong against him.