Department of redundancy department

In the past I have been accused of being judgmental, or projecting my values upon others. In fact, a few weeks ago a commenter took issue with my suggestion that those who aren’t ready to put their rooting interests in real sports aside to embrace owning stars on rival teams shouldn’t even bother inserting themselves into serious fantasy baseball leagues.

I freely admit that I can be very prescriptive as to what should and should not be; in fact this behavior is not, by any means, limited to the subject of fantasy sports. But, since I’m not here to write about anything else, I’m going to subject you to all to another prescriptive fantasy opinion. When I write here, I usually presume that the preamble disclaiming my opinions as just that – opinions – is simply taken for granted, but every now and then, I guess it’s important to explicitly state that disclaimer.

With that out of the way, I’d like to ask something of a rhetorical question regarding what I feel to be questionable decisions in league construction, particularly when it comes to delineating scoring categories. I know many of us like to experiment with tweaks on the traditional 5×5 categorical set-up. I, too, sometimes play in leagues that are 6×6, or those that substitute OPS for AVG, net SBs for raw SBs, or K/9 or K/BB for raw Ks.

However, there are certain tweaks that I do not understand. I’ve had prospective commissioners tell me that they intended to expand the traditional categories to make inclusions such as a triples category, and I’ve seen leagues that include both average and OBP as individual categories. These tweaks seem senseless to me. There are two main things that I suggest commissioners avoid when setting up scoring categories in their leagues.

First, I would suggest you avoid selecting arbitrary and incomplete niche categories. For example, triples may seem an enticing category add because triples are exciting to watch and their supply is highly limited (I’d actually argue that they are too limited, making them a bad category choice as their mere inclusion tilts the value of some players so drastically). But, the simple question is, why triples? Triples are simply one variation of base hit. So, why choose to privilege their importance or relevance above singles or doubles? The inclusion of a category like triples is not only arbitrary (granted, all category choices are arbitrary to some degree), but they represent only a small fraction of a larger group of like events.

Now, I’m sure some of you are just waiting to lash out, “what about home runs?” Doesn’t the same logic by which I criticize triples also apply to homers, too? Well, not exactly. Home runs have a few things going for them that singles, doubles and triples do not. For one, they are a part of the standard, traditional scoring system. And, yes, their inclusion there is, or rather was somewhat arbitrary, but we mustn’t let the notion of arbitrariness devolve into a slippery slope that undermines the category selection process. Home runs have a well-established place in fantasy baseball scoring and to treat the initial arbitrary nature of that decision as akin to the arbitrary nature of including triples as a category while excluding doubles is disingenuous. I’m just going to hope we can agree on that.

Second, the home run is the pinnacle of offensive achievement and it is defensive independent, which makes it a per se wise choice for inclusion as an offensive statistic. I normally frown on the inferior double counting of stats (I’ll get to this in a minute), but the home run is a unique animal. Seeing as how R and RBI are also traditional and common categories, the home run is the most pure way of accumulating offensive production; it is a hit, a guaranteed run, and a guaranteed run batted in at the same time. It is complete and pure. Unlike the double or triple, the home run is not just one of several different kinds of base hits.

The other dynamic I think prudent to avoid when setting up scoring categories is what I’m going to call inferior redundancy. I think it’s a good idea to try to encompass crediting as many offensive achievements as possible by using stats that reflect different kinds of success and distinguish between incremental successes. In the same vein, I think it is a poor idea to choose to include two or more categories when one is basically duplicative of the other(s) and more profound.

For example, I can understand choosing to use AVG, OBP, SLG, or OPS individually as an offensive stat. However, I don’t think it is sensible to use AVG in combination with OBP, as both stats are essentially aiming to measure how successful a player is at reaching base (in the case of OBP that’s what it does, in the case of AVG that’s how it is commonly misapplied). They’re more similar than they are different.

I also don’t think it is makes sense to use both AVG and SLG, as AVG measures how often a player succeeds at getting on base via batted ball, while SLG does as well, but differentiates between the value of different kinds of hits.

To me, it seems most logical to use OPS alone (or wOBA or something else that counts walks and credits extra base hits) if you are just going to use one rate category. If you need an extra category to counteract an extra category on the pitching side (I hope nobody is playing with an uneven amount of pitching and offensive categories), I think the most sensible choices are either OBP and SLG, which simply breaks the single best choice into its components, or AVG and OPS, which includes the simplest and most encompassing choices and utilizes an original, traditional stat thereby invoking the damage control argument of, “yes, this system is flawed, but I’m not introducing new flaws into the system.”

I hope I’ve been sufficiently diplomatic in stating these preferences. I have given a fair amount of thought to them and try to adhere to an internal logical consistency when I determine how I think things “should” be done. I have been known to decline to play in leagues because I think the set-up is flawed and I think everybody should be willing to do the same. If you enter a league thinking that the set up is flawed, it will negatively impact the experience you have participating.

It is entirely possible to create novel league set-ups and scoring systems, and I’m all for innovation, but league set-ups need an internal logic and a consistent vision in terms of the real life events it wants to reward. If you are tempted to create your own system, please give it thought and kick the tires before rolling it out. You don’t have to reach the same exact conclusions that I do, but you owe the effort of vetting it to all your prospective leaguemates.


12 Comments
Oldest
Newest Most Voted
Inline Feedbacks
View all comments
Don G
13 years ago

Homeruns are defense independent, but not always park independent. We get points for “lucky” or “barely over” homeruns, so why not get points for triples or doubles? Who’s on base and who scores is not the fault of the hitter. If he misses a homerun by inches in Petco, why not let him get points for the double?

We use 18 categories. Yes, really. It’s just because it’s fun and we like it. I admit we’re redundant in using AVG and OBP. We use Net Steals instead of SB.
R-1B-2B-3B-HR-RBI-NS-AVG-OBP
QS-CG-SHO-K-SV-HLD-GIDP-ERA-WHIP

We do evaluate our categories each year and change them if we think it’ll improve the scoring.

Mikey
13 years ago

I have used a 4×4 that I think works fairly well…

TB, R, RBI, SB
W, ERA, WHIP, K

Gives a little boost to the well rounded hitters.
Creates some equal ground for pitchers… why use SV but not HLD.

And requirements to qualify for 1st, 2nd, 3rd is you must be between 1400-1500IP (1458 is the AVG of innings played) and 146GP at each position (90% of all games played).

John K
13 years ago

Good post – I definitely think it’s more fun to play in leagues that think a bit harder about which categories to use.  For instance, after switching to QS, I can’t imagine going back to a W league.

One thing we did when deciding which categories to (not) add to our 6×6 system was a quick correlation matrix.  I think this could have helped you make your above point more frocefully since it is basically what you were arguing in words.

Also, awesome title

Chris
13 years ago

In a league I’m rebuilding we use salaries and FA auctions and got flak from one of my managers over instituting OPS in the league (was a standard 5×5, now a 6×6 with OPS and HLD). He was complaining that now we’re paying twice for home runs and hits.

My argument is that in a league trying to closely sim the MLB experience we should be paying for guys who get on base and guys that have gap to gap power, not just guys who hit for high AVG and HR.

Hearing you say that AVG and OPS has helped ease my concerns over potential redundancy without having to completely overhaul the system that we currently use.

The Baltimoron
13 years ago

Solution: sub OBP for AVG and XBHs for HRs. 

The flaw with AVG is that it ignores walks.  Isn’t the point of that stat to determine how successful a hitter is in getting on base?  Then why use AVG?  I think we all have already switched over to OBP as the best indicator of that specific tool (in short, on-base percentage is the best way to measure how often a player…gets on base.  Duh).  So it’s also what we should use in fantasy.

The next sub is to swap out homers for extra base hits.  The impetus for adding, say, triples, is that they and doubles are ignored in standard scoring and are clearly more valuable than a simple single.  However, while everything you say about triples is true, and they are not worthy of their own category, that isn’t to say they should be ignored.  Using XBHs gives recognition not only to power hitters, but speedsters as well.  And the types of players who rack up doubles can be anyone from Albert Pujols to Brian Roberts to Jason Werth to Adrian Beltre—not just one type or one position, but any player who can hit and get into scoring position.  Now, you might say this devalues homers, but I disagree for the reasons you listed above: HRs are automatically a run and at least one RBI, this still making them more valuable than 2Bs or 3Bs.  I believe that extra base hits gives the right balance for all elite players while also considering more relevant data for measuring a player’s worth.

Finally, I’d like to rally against using SLG or OPS as a category.  Just as triples are too limited, these categories are too broad and encroach on too many areas to be used in roto.  They may be amongst the best SINGLE category to measure a hitter, but they are too comprehensive for fantasy purposes, IMHO.

As far as pitching: there is no need to mess with the standard 5 categories.  Wins belong because it’s a stat that both pitchers and managers play towards.  We’ve all seen pitchers gut it out to leave with the lead, to outduel their opponent, and be left in when struggling to try and get the W.  I’ve NEVER seen a SP or manager make a decision about a pitching change based on getting a guy a QS.  Also, I think that the best pitchers step it up against elite competition on the other side.  If the other guy is only going to allow one run, you better pitch a shutout.  It also, like, means your team WON THE GAME.  Saves are also clearly a stat that both pitchers and managers factor into strategy (too much so, it’s often argued).  And I like K’s as a counting stat over a ratio as we already have ERA and WHIP, and elite RPs should have better rates than starters.  Using a K rate then gives way too much importance to relievers over their starting counterparts.

The Baltimoron
13 years ago

@ Don G—I think those pitching cats are pretty thin, with CG, SHO, and HLDs.  Not enough of the first two, and holds are situational, lucky, vary on a yearly basis, and don’t really reflect pitching dominance.  The trick has always been finding pitching stats that are on par with the hitting cats when trying to expand over 5×5.

@ Mikey—don’t see much reason for drafting RPs in that system.

@ Chris—wouldn’t using my settings solve the problem of “paying for guys who get on base and guys that have gap to gap power, not just guys who hit for high AVG and HR”?  If you want to value on base percentage and extra base hits, why not make those your categories?  Plus, no double dipping, and a cleaner league configuration.

And yes, I realize that my argument is as impassioned and pointless as trying to convert everyone’s religion to my own, and I’ll admit to being a zealot, but you never know—sometimes people do convert…

Don G
13 years ago

@ Baltimoron – I know what you’re saying with the pitching. I’m not really arguing that our cats are better. We just use them because it’s fun. I like seeing my guy hit a triple on Sunday to get that point back…and so on (it’s a H2H league). A few guys in the league just ignore Holds. What do you mean by pitching cats that are “on par” with hitting? Can you give an example?

The Baltimoron
13 years ago

@ Don G—Hey, I like fun categories too (my league is actually 6×6 H2H with Cycles and No Hitters added) and if y’all dig it, I certainly don’t see a problem.  I’m just advocating what I think is the best default system.

As far as the “on par” statement, I meant worthy of being their own scoring category.  There are an innumerable amount of stats kept, but we don’t use them all right?  Why include GIDP, but not Balks or Wild Pitches?  To that end, you have to ask: what makes a good category?  I think it’s the question that the article is getting at.  To me, a good fantasy baseball category should:

1)measure a real baseball tool
2)measure it as accurately as possible
3)walk the line between being too specific or too broad
4)have some relevance to real game strategy
5)be exhibited by a preponderance of players
6)factor in positional value

I would argue that, in baseball, the offensive tools you’d want to measure are: contact, power, speed, baserunning, and batting eye. So, what metrics measure them best using the above criteria?  AVG obviously accounts for contact, but ignores strike zone discipline.  To compensate, you could either add walks and/or strikeouts, but then you run into the problem of having to add more pitching categories for balance, and of trying to balance a ratio (AVG) with a counting stat(s).  So, I think it’s much more elegant to use OBP in favor of AVG to factor in both hitting and strike zone discipline.  It gets to the heart of figuring out “can this dude get on base, or is he making outs?”  Power is traditionally represented by HRs, but I think that definition is too narrow and, along with RBIs, add too much value to home run hitters.  A guy hitting 50 doubles a year has power, yet is completely unrepresented in traditional 5×5.  Again, you can take your solution of including doubles and triples as separate categories, but you then need to find corresponding pitching categories, and triples are a limited stat.  You can add Total Bases, but I don’t think that’s the best measure of power, specifically.  SLG is an option, of course, but again, you’re swapping a counting stat for a ratio, which ain’t the best H2H.  So, I arrived at XBHs as the best compromise as stated above.  HRs still count more as they also automatically dip into the R/RBI cats.  I think speed and baserunning are factored into Runs and SBs (as well as the doubles and triples contingent of XBHs), thus accounting for all the tools above in what I think is the best manner.  Best of all, using these categories gives equal weight and representation to various types of hitters in a balanced, accurate, and concise manner.

For pitching, I would want to include stats that show:

1) Can they dominate hitters?
2) Does the team win when they pitch?
3) Do they substantially contribute to the outcome of the game?
4) Do they limit baserunners?
5) Do they limit scoring?

I think ERA works well for scoring (it is still the pitchers job to limit runs, right?), while Ks and WHIP are effective measures of dominance.  Yes, you double dip that trait, but rightfully so, as I believe dominance should be heavily weighted.  Using Ks and WHIP gives you both a counting stat and a ratio (the Senate and House of fantasy baseball scoring).  I prefer using Ws over QSs, because it accounts specifically for “did the team win when they pitch.”  For me, Quality Starts is too broad, is arbitrarily achieved, only works for starters, and I’ve never seen a pitcher or coach do anything in pursuit of gaining a QS.  I think having the ratios weighed by IPs helps determine pitchers’ contributions, as well as getting the Win and closing out the game.  I hate Holds for the reasons stated previously, as they have no bearing on wins, dominance, or even baserunners or runs allowed—they are totally situational, random, and…stupid (I’m wondering if players in your league punt Holds for the season, or just at the draft?).  I think CGs and SHOs are great measures of the above, but fail the “exhibited by a preponderance of players” criteria.  I mean, you’ll get Halladay giving value there, but otherwise, it’s scarce and a bit of a crapshoot (Pavano?  Braden?) in today’s game.  Maybe that makes them more fun to you, and I have no problem with that, but it’s just too limited for my tastes (but not limited as much as No Hitters to be considered random Easter Eggs).

Finally, I’d say that defense should be considered, as well, but fails for three reasons:

1) It’s still very difficult to measure.
2) It’s unclear how much of an effect it has on the game.
3) It would be difficult to determine the right weighting to make it matter from player to player, position to position, park to park, etc.

Damn, I was intending to keep this short…

Chris
13 years ago

@Baltimoron:

The reason for this is that the league was already established and is trying to be revived. We’re bringing in new blood and keeping old managers.

Part of the problem (that I may not have been clear enough in stating) is that we are a salary league with an arbitration scale for players with less than 6 years of MLB service time. This tends to be a confusing enough system as it is with the different pay scales based on performance, and to completely remove a category and replacing it with a new one that scales differently would mean redoing most every roster entirely, just adding OPS to the batters line only means figuring out one scale and adding to the previous sums.

Don G
13 years ago

@Baltimoron – I couldn’t disagree more about Wins. Yes, usually good pitchers will win games, but there are several excellent pitchers that don’t win games because of lousy run support and there are lots of terrible pitchers that win games. We took wins out because it doesn’t measure what your pitcher actually did. If Lincecum pitches 8 IP with 1 unearned run against Kershaw’s 8 IP, no runs, Lincecum would be saddled with a loss (or non-win). That’s, as they say, “whack”!

We chose GIDP as a way to give groundballers some credit. When we set it up, GB% or any other ratio weren’t available (not sure if they are now).

With Holds, some punt them for the year, some just punt in the draft. Last year, I drafted Thornton and Bard early and basically won Holds each week (until Thornton started morphing into Bobby Jenks at the end of the year).

I like XBHs, and if we change up the scoring after next year, I may suggest that.

Just another thing to add to your argument – categories that are at the mercy of the scorer should be avoided. I realize XBHs are sometimes (was it a double, or a single-advancing-to-second-on-throw?), but it’s usually correct. We used to use Errors, then we switched to Fld%. We ended up taking them out because it’s really just a judgment call by the scorer/officials.

Nate
13 years ago

Now why do you use xBH instead of TB?  TB would take into consideration that 3B are more valuable than 2B… and HR more than either… sure there’s “double counting” the singles in OBP and TB, but w/ xBH you are already double counting… The downside, I guess, is that power hitters, once again, have the upper hand… thoughts?

The Baltimoron
13 years ago

@Nate—Yes, Total Bases makes sense, too, but it does seem to skew things towards HR hitters and less for the speedy guys.  HRs already get a run and at least an RBI, so they are still more valuable in this setup.  Honestly, the original goal was to factor in all XBHs, so that was simply subbed for homers.  Perhaps most importantly, calculating XBHs is easier without the weighting required for TBs!