|How to view standings objectively|
|Theory and Strategy - Platinum|
|Written by Todd Zola|
|Sunday, 10 June 2012 11:06|
It's a marathon, not a sprint. Patience is a virtue. It's still early. … You've heard them all, no doubt. But here's a relevant question: At what point is it not early?
That's what I'll examine today, when it's time to objectively take stock of your team and decide if, say, Felix Hernandez is expendable because you now have Christian Friedrich, or whether Michael Bourn is helping you as much as you think. Most importantly, I aim to show you why it's never too late to gain or lose points in batting average, ERA and WHIP, which is contrary to popular belief.
To do this, we must put Rotisserie league standings under the microscope. Each category will be dissected, and some of the nuances and subtleties with respect to game theory will be elucidated. Is it too early? What's the best way to objectively assess and improve our squads at this point in the season? We'll look at that with this underlying theme: Understanding the ins and outs of the standings can aid in your endeavor to maximize Rotisserie points.
Apples to apples, oranges to oranges
The vehicle that we'll use for our examination is what I call "normalized standings." What ensues is the process to generate Rotisserie league standings so that each category can be compared to another in a simple eyeball manner. To eliminate as many outside variables as possible, year-end standings for the National Fantasy Baseball Championships (NFBC) will be used. Each NFBC league is a 15-team mixed league with a snake draft. Each league's draft occurs simultaneously. The sample for each season is at least 20 leagues. There's no trading in the NFBC, but there is weekly free-agent bidding with the ability to activate players from a reserve list. The fact that there is no trading, and all drafts occurred at the same time so the information available with respect to injuries and job battles was consistent helps minimize selection bias.
The key to normalized standings is converting each category to a unit that allows for direct comparison. The way we'll do this is by having the total units in each category be the same number. We'll choose 1,000. Even though the ratio categories (batting average, ERA and WHIP) are all based on different at-bats and innings pitched, they will be treated as counting categories. Because ERA and WHIP are scored with the lowest being the best, the reciprocal of the ERA and WHIP will be used for this calculation.
I gathered data from the 2004-2009 seasons, excluding 2007 (I lost that data on a misplaced jump drive). The 2010 standings were not available, and the 2011 data was removed the NFBC site before I had an opportunity to archive it. But as you will soon see, the trends are strong enough to draw tangible conclusions from the data that is available for this study. As a means to present the data in a more efficient manner, the first-place total, the seventh-place total and the 15th-place totals are shown. The trends are representative of all the standings places.
Here are the normalized standings for the five years' worth of data I collected:
As you see, we're not quite there yet, but sure enough, the top-to-bottom distribution is narrowing, and within a week or two, the standings will have stabilized so that if you prorate them to the end of the season, they would look normal. That is, each team would have accrued the usual percentage of stats within each category.
The tables are pretty, but what do they mean?
As a quick aside, it's important for you to interpret the data in any means you see fit and make them applicable to your specific league format. I quickly learned after attending Chemistry Group meetings in graduate school that there can be multiple interpretations of the same data. Sometimes it takes follow-up experiments to discern right from wrong. Other times, there actually are multiple ways to explain the same set of data.
With that as a backdrop, what the pretty chart above fails to tell you is that if you are not in first place by the end of May, you might as well start ranking your running backs for fantasy football. But the tables can give you a realistic snapshot of where you stand in each category, and how much you might need to improve to earn more Roto points and ascend in the standings. After all, trades and waiver pickups are not about player value or ranking; they're about the potential to gain points. Standings places are the true currency of Rotisserie baseball, not a player's dollar value or rating.
With that, let's have a brief discussion on each category, with telltale signs you should look for to objectively assess where you stand at this point of the season. It needs to be stressed that this exercise entails looking at average standings. By the nature of the beast, the gap between consecutive standings places has been smoothed out. This is not to say the distribution is linear within each category, because it's not. However, each individual league likely had a cluster of tightly grouped stats with large gaps between adjacent teams elsewhere. Ultimately, when you are strategizing, it is where you're situated within each cluster that will dictate your path to more points. There are some categories in which it's easier to gain or lose points, but just to reiterate, this is based on average standings that have been smoothed. Your mileage within your league may vary.
The big picture
As a scientist, I get a charge out of results that are unexpected or even counter-intuitive. With respect to fantasy baseball, if you are aware of the results and others are not, that presents an area in which you can gain an edge. Below is such an example. This is a consolidation of data that has already been shared, but re-organized in a means to highlight a very interesting finding. Using the average normalized standings, the first place, last place and difference is presented, from smallest to largest difference:
It makes sense that the stolen bases and saves categories have the largest top-to-bottom difference. Perhaps that's because, inevitably, in every league someone tries to punt steals or saves, while someone else loads up on them.
What strikes me as intriguing is the three ratio categories that are the most tightly bunched in terms of top-to-bottom distribution. This is an advantage of using normalized standings, as one would be hard-pressed to observe this simply by looking at the BA, ERA and WHIP numbers.
Before I offer what I believe is an actionable repercussion of this factoid, let me assuage those that are concerned that the above is an artifact of teams punting the singular categories, or poor teams giving up by running a regression on the entire category. I will do this by showing the difference between No. 1 in the standings and No. 15, then incrementally removing a team from either end to illustrate the relative movement necessary to gain or lose points:
ERA and runs flip-flop, but the larger point stands: In general, the ratio categories are more tightly bunched than those of the counting variety. My immediate reaction is to question the theory that after a certain point in the season. It's impossible to gain or lose points in the ratio categories. I know that as at-bats and innings mount, the amount your ratio can move does lessen, but the fact is, it does not have to move that much (in many cases) to gain or lose points. Keep in mind the above is based on average standings. The distance between adjacent places will not always be the above linear regression in the table; there will be clusters where the gap is smaller. The category is already more bunched than others, which suggests there will be hot spots within the ratio categories where it's quite possible to gain valuable points if you are fortunate enough to be situated in the right spot, especially because, while you will not lose counting stats, ratios can worsen. Think of it this way: You have never gained a point in homers on a night where none of your batters went deep, but you probably have gained ERA and WHIP points on a night in which your staff was idle due to a competitor starting, let's say, Derek Holland for a juicy home matchup against the light-hitting Seattle Mariners. Not to belabor this, but I have in fact-tracked the movement within the 10 categories in September, and there is indeed more action in the ratios than the counting stats.
We'll conclude this discussion with a survey of each category, pinpointing what to look for to objectively gauge where you stand. The average category numbers from the NFBC leagues comprising the study will be presented as well:
You won't see them here, but the first-place teams in the category the past couple seasons have been in the 270 range which is not a surprise considering runs and homers have trended downward of late. The takeaway: A 35-homer hitter is more valuable today than he was five years ago. Even though the total number of home runs hit has decreased, in terms of Roto standings, the relative distribution within the category is the same, so those same 35 homers can gain you an extra point or two in the standings.
Perhaps the most relevant aspect of the home run category is that, historically, league champions fare better in homers than any of the other hitting categories. Keep in mind, all the data here is calculated by averaging the numbers of the first-place team in each category and then on down the line, as opposed to averaging the number of homers the league champion scored, then the second-place finisher, etc. I also crunch this data based on NFBC leagues, and homers has always been the one in which champs score the highest. Intuitively, this makes sense, since RBI and runs come along with long balls. The bottom line is if you are still struggling in the home run category, while you can still have a couple of slow starters, such as Adrian Gonzalez and Rickie Weeks, get their power stroke back, it may be time to address the deficiency via trade or free agency.
As you might expect, batting average has dropped about 10 points across the board since 2009. The key here is perception. Those of us that have been playing this game for a long time have a sixth sense for the value of a player just by hearing or seeing their stats. Depending on the size of the league, there's no way one would roster someone hitting .240, for example, regardless of their other production, but .250 is okay as long as it came with some pop and/or speed. Well, folks, .240 is the new .250, as the kids say. It is tough, but some of us have to make that mental adjustment when we see a stat line for an available free agent.
My favorite thing to do with respect to batting average is compare where I am with respect to runs and RBIs. If my category points in average are lower than in runs and RBI, I am optimistic that when my guys start getting more hits, I will do even better in the other categories. However, if my average is high but my other production is low, I am a bit concerned, since the way to address this has to be by increasing homers. BABIP correction alone is not going to get it done.
We'll put these categories together since they are somewhat related and correlate pretty well. More often than not, your placement in runs and RBI are only off by a few standings places. If they are separated by a wide margin, you need to take a look at the construction of your team. Chances are you have a plethora of hitters at the bottom of their lineups or playing for squads with below-average offenses.
Sometimes we take RBIs and especially runs for granted and only focus on batting average, homers and steals, assuming the rest will come. And most of the time, it does. Having the large gap in these categories is nothing to panic about, but it should instigate a peek at your team to see if there are any players with a vacuous batting average that you can parlay into more counting stats. Truth be told, the best way to keep the runs and RBIs in check is to maximize at-bats. In daily leagues, this means paying extra attention on Monday and Thursday when you may need to make a move or two to account for an off day. In weekly leagues, this means being on top of all playing-time situations for your fringe players. Sometimes it's easy to miss when a player has fallen into a platoon. Sorting the previous week's stats by at-bats is a great way to check this out. Something else to do is be aware of the scheduled opponents of your lefty hitters. If Luke Scott is scheduled to face southpaws CC Sabathia, Andy Pettitte, Jon Lester and Felix Doubront in a given week, you should seek a replacement since that's four games he (and other left-handed batters) might not start that week.
Earlier, stolen bases were referred to as a singular category. This is because a small percentage of players contribute a large portion of pilfers within the category. The singular nature of the category means it is less correlated to how well you are faring in the other categories. A high average does not mean a lot of steals. You can have a competitive runs total without speed demons. This leads to all sorts of game theory from punting steals to getting Michael Bourn and a little bit of help to over-drafting steals, so you can deal them for help later in the season. Keeping in mind that everything depends on your particular locale within the category, the regression data above suggests, on a relative basis, it takes more steals to move up in a category. In addition, based on the raw data trend, as power has waned, steals have increased, so the 30 swipes you got from Derek Jeter a few years ago were worth more than the 30 you now get from Jason Kipnis. But to reiterate, everything comes down to where you are in your standings and how many points Jose Reyes, for instance, may net you versus how much you lose elsewhere.
There really isn't too much to say here. While good pitchers on teams with good offenses and bullpens have a better chance for wins, there will always be surprises. Perhaps the most pertinent lesson from the wins data is that, after steals and saves, the most production is necessary to move up in the category. So not only are wins a crapshoot, even if you luck into a few, it might not be enough to jump a competitor in the standings, depending on where you sit.
Like runs and RBIs, ERA and WHIP correlate pretty closely. But unlike their hitting brethren, there is a definite cause and effect with the pitching duo. Of the two, WHIP is more real and less luck-based, so it's most often the entity that best describes where you are at. As the season progresses, more often than not, regression kicks in and your ERA moves toward your WHIP. This is true in either direction. The main exception is when you have a disproportionate number of fly ball or ground ball pitchers on your staff. Generally, extreme fly ball guys give up fewer hits since fly balls that stay in the yard can be caught, resulting in a low WHIP, but those that do clear the fence elevate their ERA. So a preponderance of fly ball pitchers may result in your ERA earning fewer points than your WHIP. Ground ball hurlers tend to give up more hits but fewer runs, as they also induce more double plays and surrender fewer homers. In this instance, your ERA might score more points than your WHIP. If your ERA is truly out of sync with your WHIP and the correction is favorable, more wins may be coming your way, and vice versa.
While this may seem to be contradicting the earlier contention that it's never too late to gain points in ERA and WHIP, we are at the point in the season where one of the most difficult decisions must be objectively considered: Are my ratios so poor that it's best that I just blow them off and only worry about wins and strikeouts? Anytime you opt to punt a category, let alone two, you are seriously reducing your margin of error, especially in keeper leagues. But sometimes doing so avails assets to help elsewhere, and the supply of poor-ratio pitchers is usually pretty high with the demand low, so the cost is favorable. There is a good chance that someone like Derek Holland, Christian Friedrich or even Ivan Nova is available in a mixed league. If you are forgoing ERA and WHIP, who cares how hitter-friendly The Ballpark in Arlington, Coors Field and the new Yankee Stadium might be. I'll take those four earned runs in six-plus innings as long as they come equipped with five punch-outs and a W. I can then deal my Felix Hernandez and Bud Norris, good pitchers on lesser teams, and load up on such guys as Matt Harrison and Joe Blanton. It's all about the win, baby.
Strikeouts is another category for which there is not a whole lot to say other than to remind you that if there is an innings or start limit in your league, the K category is really K/9, so it's best to load up on high K/9 starters and only start guys like Mark Buehrle and Jason Vargas if you have no other choice.
Even the Hubble telescope isn't powerful enough to fully treat the saves category. You know the deal here. Perhaps the only real advice that can be shared when it comes to saves is the damage a poor closer can do to your ERA and WHIP which is now further evident based on how tightly bunched the ratio categories usually sit.
That wraps up our survey of the categories, as well as dissecting the standings in general. Remember, the single most important piece of information as you scheme to gain more points is to do the math based on where you are within each category. But hopefully the above analysis shed a little light on typical Roto standings so that effort has been made a little easier.
|Last Updated on Sunday, 10 June 2012 11:27|