The Long Game: Clowns To the Left ...

Way back at the end of April I took a look at some ways to judge whether your team was truly a contender or not, even if it got off to a less-than-ideal start. Well, now it's three months later, and some of you out there are probably still on the fringes of contention, wondering whether you should go for it all or play for next year.

This late in the season, being mired in the middle of the standings can be an ugly spot to find yourself in. Depending on how active your league's trade market has been, most of the big prizes may have already been dealt. In AL- and NL-only leagues reinforcements could be on their way thanks to MLB's trade deadline, but there's no guarantee the players that get added to your talent pool will actually be useful to you.

What I'm going to share with you now is a super-duper-top-secret method I've been using for many years which provides me (and, now, you too) with a quick-and-dirty snapshot of my realistic chances at winning, or at least finishing in the money. This wonder tool can provide you with a far better understanding of your true odds at making a push in the standings, which can spell the difference between accurately-targeted player additions that give you the boost you need and wastes of time that squander your precious keeper assets for nothing.

What's the name of this magic device? Why, it's none other than your old friend from Statistics 101, the standard deviation!

If you aren't familiar with this amazing gizmo (warning: lay person's explanation of a mathematical concept approaching) standard deviation is a statistical term that essentially refers to how tightly grouped a set of data points are around their average. The tighter the grouping, the smaller the standard deviation (in other words, the less those data points deviate from the norm). This is based on what's known in probability theory as a normal distribution, which says that the majority of the data points drawn from the same source are going to be clustered around the average. Roughly two-thirds of the data set will be within one standard deviation of the average, while just over 95% will be within two standard deviations. The traditional way of showing this visually is with a bell curve. The top point of the curve is the average value of the set, and the data points themselves are plotted as either positive or negative values relative to the average. The slope of the curve illustrates the size of the standard deviation. A steep slope means a smaller deviation, as more numbers stay close to the average, while a gentle slope indicates a larger deviation and a wider spread of values.

What it all means from a statistical perspective though isn't as important as what it means from a rotisserie perspective. Every scoring category that your league uses contains a set of data points that are the season-to-date results of the teams in your league. In a 12 team league, you'll have 12 data points in home runs, 12 data points in RBI, 12 data points in wins etc. Figuring out the standard deviation of each of those sets gives you a glimpse into how tightly bunched those categories are, and also gives you a quick way of determining how likely it is that you can move up in those categories.

Is your head swimming? Don't worry. The good news is that you need almost no math skills whatsoever to figure this out for your own league. All you need is Excel or another spreadsheet program to do the work for you (see the bottom of the article for the formula).

Just to demonstrate how this magic instrument works, I'm going to use my own team in the RotoWire Staff Keeper League as an example. It's an 18-team mixed 5x5 league with a fairly tight salary cap in which I'm currently spinning my wheels in eighth place. For instance, in batting average the top team is cruising at .2724, while the 18th place team is limping along at .2366. I, sadly, am closer to the bottom than the top at .2543, good for a miserable 15th place. All the listed numbers were pulled just this past weekend, and I've assembled them in the following chart, with the standard deviations for the bulk production categories rounded to the nearest whole number:

As you can see, there's a whole lot of mediocre going on there with my squad. But being that low in so many categories just gives me more opportunities to gain standings points, right? Well, not necessarily. That's where the standard deviations come in.

Let's say, for the sake of argument, that I have a very good second half and make up an entire standard deviation's worth of ground in batting average, getting me all the way up to .2640. You know where that would put me in the standings? 13th. I'd gain a whole two points. Yeesh. But wait, it gets worse. Calling the period after the All-Star break the second half is just lazy shorthand. There were actually only about 67 games left in the season when I pulled these numbers (the average MLB team had played about 95 games), or about 40% of the season left. Factoring that into it, and I really can only reasonably expect to gain one standings point (the team two places above me is just out of reach at .2619). Double yeesh.

Other categories, though, look much more fruitful for me using this method. A three-point leap appears possible in HR, a massive five in RBI (second place is at just 551, 27 ribbies ahead of me), another three in runs, three in ERA, four in K's. What's more, the one category I'm rocking in (saves) poses very little risk. Second place is right on my heels at 61, but third is at a distant 52 and I could normally expect to lose at most one standings point if I drop one standard deviation's worth.

Now, is it possible to rise more than one standard deviation in a category? Of course! Miracles do happen. If you clear out your farm system to add every elite starting pitcher you can lay your hands on, you could see some incredible rises in ERA, WHIP, K's and wins, for instance. But it's not very likely at all, and would take an incredible set of circumstances to pull off. Remember, statistically speaking only about one in 20 data points lies more than two standard deviations away from the average. What we want to look at here is not the theoretically best possible outcomes, but merely the statistically likely positive outcomes.

Totalling it all up (gains in nine categories and a drop in saves) and the best I can reasonably expect to do this year if things go right is ... third place, from just shy of 100 standings points to the mid-120s. That's not bad given that the top six finish in the money, but it's certainly not what I was hoping for. More importantly, it's not a result that encourages me to go all-out this year and trade away a ton of keeper value for expiring and expensive contracts. I've got a solid shot at a money finish, but effectively zero chance at a title.

So, instead of pursuing players like Joey Votto (he's been dangled) to try and make up those precious HR/RBI/Runs points, I made a smaller deal to pick up Kendrys Morales in exchange for one of my three closers (Fernando Rodney). That, combined with better health and performance from players like Giancarlo Stanton, Brandon Beachy and Martin Prado, as well as further contributions from players picked up in a trade I made in June when I thought a title was still possible (Adrian Beltre, Lorenzo Cain and Zack Greinke, and yes I now feel sick about the massive amount of keeper value I paid for them), should hopefully be enough to climb that ladder.

Keep in mind, this isn't intended to be a predictive tool. Just because the standings gains are within reason doesn't mean they're necessarily within reach. You'll still need the players, and the luck, to pull it off. This method is simply a way to see what can be done, and thus give you a better perspective on how much of your keeper value stockpile you want to commit to the chase.

How to calculate standard deviation is Excel:

1) Copy the data set for one of your league's categories into a column of cells.
2) Select a different cell to display the standard deviation
3) Type '=STDEV(data set range)' into that cell
4) Repeat for each category

'Data set range' refers to the range of cells containing the data. So if you put the HR totals for a 12 team league right up in the top corner of a sheet, the data set range would be A1:A12. You can also just highlight the column of data as you are typing the equation, and Excel will figure it out for you.

Once you have the standard deviation for a category figured out, you'll want to use about two-thirds of it to reflect your reasonably possible gains in that category. That's because about 60% of the season is already in the books, and 40% divided by 60% is two-thirds. (You don't need to be exact, so don't worry too much about getting it down to the decimal.)