Several years ago, I introduced the concept of Category Efficiency Rankings (CER). The notion didn’t gain traction so it was abandoned. The idea involved massaging a conventional valuation set to render it more practical.
Fueling the move away from CER was a deeper understanding winning was less about pinpoint valuation and more about game theory and roster construction. Valuation shouldn’t be thought of in absolute, but rather in relative terms. Introducing a means of fine-tuning valuation simultaneous with preaching a more abstract drafting approach was confusing. Why do I have to think if CER does it for me?
Before delving into why it’s time to bring CER out of hibernation, let’s review valuation theory in general terms. Valuation is the process whereby a player projection is assigned a dollar value, ostensibly to be used in auctions, but also to proxy as a cheat sheet for drafts via ranking players by descending value.
There are three primary systems. We employ the Percentage Value Method (PVM). PVM distributes dollars in proportion to the players’ contributions to the respective categories. The most popular process is Standings Gained Points (SGP). Here, money is assigned relative to how each player helps a team earn points in each category. The last procedure uses standard deviations (SD). The player’s expectation is scored based on how many standard deviations it is from the mean of each category, then budget is distributed as a percentage of the player’s summed standard deviations across all categories to those of the field.
Regardless of the method, valuation is flawed. Plus, an output is only as good as its input, and the error bar associated with player projections is estimated around 30 percent. The dollar value is not only a function of the player’s projections, but also encompasses that of replacement level players and the cumulative contributions of the draft-worthy field. It’s plausible to nail a player’s projection but to miss on the dollar value if the other factors are off. As mentioned, my preference is for PVM. SGP is intrinsically flawed while SD lacks a theoretical basis.
The flaws with SGP pertain to the non-linear nature of each category. If it takes 5 units of a stat to gain a point in that category based on average standings and the player projects to producing 20 units, they are assigned four SGP (20/5). However, when you add up the SGP for each category, they’re all different. All categories count the same, therefore an equal number of SGP should correspond to each.
The discrepancy emanates from the uneven, top to bottom distribution in the categories. The lowest team gets one point. To earn more, other teams need to eclipse this barrier. Let’s define barrier SGP as the last place total divided by the units necessary to gain a point in that category. If stats were evenly distributed, the barrier SGP would be the same across the board, but they’re not. Typically, the barrier SGP for stolen bases is much lower than the other hitting categories (the same for saves in pitching) skewing the balance. While this is a flaw with the system, as will be explained later, it serendipitously acts in a positive manner. Still, I have an issue endorsing a method that’s right for the wrong reasons.
The SD method exhibits the same downfall as SGP; the sum of the SD per category aren’t consistent. In addition, while the SGP method has an elegance in its genesis, there’s no theoretical basis between standard deviations and the extent the player helps a team succeed.
PVM is far from perfect, but at least the theoretical foundation and math make sense. Other than the flaws inherent to all methods, the primary shortcoming of PVM is while value is allocated proportionately, spending the same amount in each category doesn’t result in equal points in each. The reason is the unequal distribution of stats within the categories.
They say a picture is worth a thousand words so let’s use an image to illustrate the above. Data from the 2016 National Fantasy Baseball Championship Main Event will be employed. These are 15-team, mixed leagues with no trading so the player pool is as homogenous as possible. Results from 30 leagues were used to determine average standings. That is, the average total was calculated per category for each standings place.
The key is converting average standings to normalized standings. The entails assigning each category an equal number of units, we’ll use 1000. The average stats at each standings point are scaled so the sum of the 15 spots is 1000 for each. Here’s the plot:
As suggested, the last place team in stolen bases has fewer units relative to the other categories. As one may intuit, runs and RBI track closely. Homers have the second lowest last-place barrier but it’s significantly higher than steals. Most noteworthy is the flatness of the batting average distribution. This will be of utmost importance when we get to CER adjustments.
Let’s save another 1000 words to graphically display how spending the same amount in each category yields a different number of points. Here, the budget allotted for each hitting category is computed based on conventional valuation methods. The standard budget is $260. While this differs by league, on the average, 69 percent of the budget is spent on hitting. These leagues have 15 teams with five categories so the amount dedicated per hitting category is $260 x 15 x .69 / 5 = $538.20.
Similar to categorical stats being normalized to 1000 units, they can be normalize to $538.20 showing how many points should be earned by spending a specific amount:
Keep in mind we’re all looking for a positive return on our investment, so if we spend, say, $36 in a category we’re hoping to accrue over $40 worth of points. Still, either you need to spend a lot more or get a huge profit in steals to finish in the upper third of the standings. This is relevant since league champions generally amass 115-125 points, an average of 11.5 to 12.5 in each category.
The $36 above wasn’t arbitrary. Distilled down to the team level, equal spending per category amounts to $35.88 each. Not-so-coincidentally, that lands at the midpoint in the standings per category, hence the obvious need to derive profit to win. For what it’s worth, it usually takes $320 worth of stats to drive down victory lane
Relating this to CER, doesn’t it make sense to allocate more budget to homers? Adding $3 to homers yields 13 points while subtracting $3 from steals renders 6 points for a total of 19. Equal spending returns 16 points. Again, profit is expected regardless of the amount spent. But even that favors funneling budget away from steals and to homers. The landing in steals is just past where points between adjacent teams is greater. On the other end, just a little profit in homers earns 14 or 15 points, along with contributing to runs, RBI and even batting average, something relevant that hasn’t been broached yet.
Speaking of batting average, look at how little money separates first and last place, barely $3. Plus, batting average is the most unpredictable category as there’s more variance. Shunting budget from average may not even sacrifice points if one or two hitters get a little lucky with batting average on balls in play. Truth be told, aiming for a low average in lieu of more counting stats is a common ploy. The above chart lends credence to the tactic.
There are two more reasons favoring drafting homers in lieu of steals. The first is the way league champions fared across the hitting categories. Here’s the average number of points the 30 league winners totaled in each of the hitting categories:
Victors scored highest in runs, homers and RBI, a full standing place higher than steals and two more than batting average. This is in perfect sync with the conclusion drawn from the charts above; spend more on homers, take away from steals and batting average.
The other justification takes advantage of what’s available in season. Traditionally, more stolen bases than homers are added via free agents. One way to show this is comparing drafted stats to season-ending numbers. Using the NFBC Main Event, the stats derived from treating opening-day rosters as a draft-and-hold, then comparing those totals to actual final numbers, historically 90 percent of homers are acquired on draft day, in contrast to 85 percent of swipes. It’s not much of a difference, but is sufficient rationale to focus more on homers at the draft table.
Pulling things together, there are three viable reasons to devote more draft-day attention to homers at the expense of steals:
Those with extensive experience can likely assemble a quality squad without CER adjustments. But why not get a nudge from more practical rankings?
Before addressing that question, let’s circle back to the contention that SGP, and to a lesser extent SD, are right for the wrong reasons. Here’s the 2016 end-of-season earnings for the top hitters per each system:
The pair to focus on are Jonathan Villar and Nolan Arenado. Note how much higher PVM places the stolen base leader Villar while lagging on Arenado, who swiped only two to complement his 41 homers. Serendipitously, SGP accounts for the three bullet points above.
Now consider that SGP is the most popular valuation process so most of your league-mates will be guided by SGP ranks. Experienced players intuitively know to go the extra buck on sluggers like Arenado and be careful with speedsters like Villar or their team will suffer a paucity of power in lieu of a surplus of speed.
Until recently, I opted to direct writing energy towards this Zen style of drafting, rooted in numbers and analysis but with an overall awareness of team construct. While this is still largely the manner I’ll draft, there’s a tangible benefit to utilizing CER. Hence, the decision to relaunch the concept, in hopes this time it germinates.
Two obvious questions are now on the table. What should the adjusted category weights be and what about leagues of different size and format? Let’s work backwards.
All the described trends transcend standard 5x5 leagues. The categorical distributions are slightly different, but in the abstract, the same three reasons supporting CER for 15-team mixed leagues hold true in all leagues.
That leaves the big question, how should the hitting budget be allocated for most efficient spending? Unfortunately, there isn’t a one-size-fits-all answer. Plus, remember values and ranks are still just a guide. Reading the room to optimize roster construction is still paramount.
At the beginning, it was stated that valuation shouldn’t be thought of in absolute but rather relative terms. It’s not that a $25 player is worth $25 and a $24 player is worth $24. It’s that’s in a vacuum, the $25 guy contributes a little more than the $24 guy. Who knows, maybe the contributions of the $24 player are more useful based on your team construct. Still, the better the relative ranks, the better your basis for making decisions.
Here’s an example of a logical category adjustment based on the data presented for the NFBC Main Event content. It was already demonstrated that taking $3 from steals and giving to homers results in three more points so let’s start there. The flatness of the batting average curve screams to borrow from Peter to pay Paul, so let’s take $2 and give a buck each to runs and RBI. Lastly, to account for the second and third justifications for CER, let’s grab another buck from steals and award it to homers. Keeping in mind the starting point is $35.88 for each, the final dollars per category are:
My value calculator allows customization of the category weights. The default is 1 for all five, which was used to generate the PVM earnings presented earlier. Converting the adjusted budget distribution to category weight yields:
Plugging those coefficients into the value calculator renders CER earnings, displayed with those already shared:
Villar is still higher with CER but he deserves it, in a vacuum steals are extremely valuable. Sure, we can empirically adjust the category coefficients to match SGP, but why not just use SGP? From where I sit, being guided by the better PVM process, then monitoring team construct is the best of both worlds. The impact of Villar, and other speedsters, is softened, but not mitigated. This seems fair.
With respect to Arenado, and other sluggers, their value is poked up to almost match SGP and SD. Understanding the imperfection of projections and valuation theory while keeping in mind we’re talking about relative and not absolute rankings, the CER adjustment is plenty convincing to go the extra buck or two so you aren’t devoid of ample power. The kicker is if you’re bidding against others using PVM, you may split the difference and not have to bid the full CER increase.
Well friends, there you have it, the case for CER, part objective, part subjective. While some may be wary of the subjectivity, that’s the beauty of it. It’s not a one-size-fits-all entity. Some need more of a push than others. Different leagues have different characteristics, especially those of the long-standing variety with established trends. Combining the advantage CER offers along with your expertise is a surefire recipe for success. Let’s swap victory stories in the winner’s circle!
Can you believe it? We’re about to embark on our 21st season on this Interweb thing. Who knew, way back in March of 1997, what the next twenty years would bring. For me, it paved the transition from a career as a peptide chemist to making my living as a fantasy writer. We’ll have plenty of time to reminisce as we celebrate our 20th anniversary in the spring. Of immediate focus is…
Platinum 2017 will be launching on Tuesday, November 15!
Once again, we’re the FIRST source of projections, rankings and dollar values specifically tailored to fantasy baseball enthusiasts, like yourself. The initial drop will be just projections. As always, we’ll roll out all our additional content throughout the off-season. This includes Lawr Michaels unique Top-250 Minor League rankings, Perry Van Hook’s minor league rankings focusing on the 2017 season and of course my player profiles and all the other goodies to aid in your draft prep.
New this season, I’m writing The Z Book, a collection of strategy and game theory essays. I promise you won’t find anything like it anywhere. The Z-Book will be part of your Platinum subscription as well as available for separate purchase. The ETA is February 15, giving you ample time to digest everything in time for the primary drafting season in mid to late March. Topics will include
Let’s be honest. If you’re already yearning for projections and rankings, there’s a good chance you’re preparing for the early pay-for-play drafts. Granted, there are exceptions, those that want to get going on keeper and dynasty league preparation but most early subscribers are getting ready to draft in high stakes contest. The majority of these are of the draft and hold variety, with ultra-deep rosters. As such, you’ll want projections and rankings for everyone with a chance to see action anytime in 2017.
Don’t worry, we have you covered!
We’ve gone over the prospect lists and will have a projection for any player with a modicum of a chance to play this season. Anyone missed that’s drafted in any national contest will be added to the next posted projection set. Shoot, if you’re thinking of drafting a player not projected, drop me a line and I’ll ask my little black box to do its thing.
More good news for those that play in leagues with non-standard categories. I’ll be adding new statistics to the projections, most notably quality starts for pitchers among others. I can open the floor for requests, with the understanding I may not be able to honor it. If the data is available and the projection is easily automated, I’ll gladly add it.
Why use Mastersball Platinum? Four of the past five National Fantasy Baseball Championship (NFBC) Main Event Champions are subscribers. This includes our own Greg Morgan, who writes a monthly column on the NFBC, available free the first Thursday of each month. The Mastersball staff aren’t just writers, we play too. Lawr Michaels is a two-time AL Tout Wars champion, Brian Walton has won NL Tout Wars while Zach Steinhorn is the defending champion of Mixed Auction Tout Wars. Pardon the humblebrag, I won both NL Tout Wars and Mixed LABR in 2016, along with being among the career earnings leaders in NFBC auctions, including winning the NL only league three times and co-owning an Ultimate Auction winner.
Questions? Hit me up in the comments section.
Fasten your seatbelt, 2017 promises to be our best season yet!
How to handle catcher pricing and ranking is an oft-debated topic. If positions are ignored and a positive price is assigned to exactly enough hitters necessary so everyone has a full active roster, the number of draft-worthy catchers isn’t sufficient to stock them all if positions aren’t ignored.
As such, there is a contingent of fantasy players that contend some owners are forced to pay a positive amount for players that will return a deficit. Others insist that pricing should be adjusted such that there are at minimum, ample players at each position with a positive price to comprise all the active rosters. I’m a card-carrying member -- if not the leader -- of the latter.
Mathematically, this is accomplished by setting the price of the worst drafted player at each position to the minimum cost and scaling everyone else up. Our Platinum archives store a plethora of essays on this topic.
Good news, gang. As part of an impending site facelift, said essays along with a treasure trove of additional strategy pieces will soon be brought out from behind the firewall and will be available for free in our new MastersVault.
Let’s take a look at what’s actually happened the previous three seasons in terms of catcher earnings at the end of the season. Dollar values from 15-team mixed leagues, standard 5x5 scoring with 23-man (two-catcher) rosters will be calculated – both forcing positions and ignoring positions. The top earning along with number 30 (worst positively valued) backstop will be presented.
Yikes! Last season’s catcher bump really spiked.
Values are reflective of the player’s contribution to the draft-worthy populace. To force positions, each player’s raw stats are adjusted by subtracting away the respective replacement level by position. This sets the worst draft-worthy player at each position to the minimum price, conventionally $1. The worse the replacement level, the bigger the price bump since fewer stats are taken away from the player.
Perhaps a look at replacement level catcher stats will help elucidate the impetus for the alarming catcher bump last year.
Sure enough, the replacement level receiver is indeed declining – most notably in batting average. There’s a cascade effect as a lower average results in fewer runs and RBI. For those unaware of valuation protocol, there is a means of converting batting average into a pseudo counting stat. Last season after this conversion, each catcher’s adjusted average along with being docked fewer runs and RBI resulted in the huge delta between forcing and ignoring position.
So what’s the early forecast for catchers based on our 2016 projections? All you need to know is, at least presently, Buster Posey is the third ranked overall player (not catcher, PLAYER) with Kyle Schwarber occupying the seventh spot.
Bad news, gang. The rest of this discussion is for Platinum subscribers only. Before you get all pissy, please realize the aforementioned site revamp is subsidized via Platinum sales. In other words, you’ll be benefiting from the site upgrade made available by the Platinum subscribers. That said, I promise the remainder of this discussion, along with a bevy of upcoming essays, is well worth the cost. This doesn’t even consider you’ll have access to the same player projections, profiles and tools utilized by four of the past six National Fantasy Baseball Championship grand prize winners. This includes our own Greg Morgan, who along with his father Dale took down the 2014 NFBC title.
I apologize for the vagueness of this addendum but we’re excited to be on the verge of unveiling a promotion where you’ll have Platinum access at a significantly lower cost. However, if you’re anxious to subscribe now and are eligible for the promotion, we’ll be happy to honor it with your payment adjusted accordingly.
Hang tight everyone. We have some exciting plans in store for the 2016 campaign.
New for the 2016 fantasy baseball season is a unique contest offered by our friends at the National Fantasy Baseball Championship (NFBC) called the Cutline Championship. A complete review of the rules can be found HERE.
In brief, the Cutline Championship is a points-based, best-ball scoring format. The leagues consist of ten teams and use standard NFBC roster requirements and position eligibilities. There will be an initial snake draft to fill 36 roster spots then a pair of in-season FAAB periods. The first is the week after the season starts where you can add up to five more players with the second in early June where you can add as many as you want to a maximum of 46 roster spots. The regular season ends right around the All-Star break where teams will be entered into the Cutline Finals, Consolation Round or have their season end. More teams will be eliminated over the next nine weeks until a Cutline Champion is crowned in early September.
What follows is a primer for those entering the inaugural Cutline Championship. Even though the discussion will focus on that contest, many of the principles transcend into other formats, so hopefully all Platinum subscribers can glean a nugget or two to help in their draft prep.
The Cutline scoring is designed so that the ranking of the players by points emulates the ranking via standard 5x5 rotisserie scoring. The hitter’s correlation coefficient is .99 while the pitcher’s is about .90.
A noteworthy difference between the Cutline and other NFBC contests is there isn’t a Friday transaction day for hitters. The scoring period for everyone runs from Monday through Sunday.
Points are awarded as follows:
For those not familiar, best-ball scoring means your optimum lineup will be determined automatically each week without you ever setting a lineup. The only team management required is the initial draft and the two in-season FAAB periods. The NFBC site does the rest.
The intelligence is designed to account for corner infield, middle infield, utility and multiple position eligibility. There’s no delineation between starting pitchers and relievers – your top nine arms each week contribute to your total, regardless of their role.
PROPER RANKING USING POINTS SCORING
As discussed, back-testing using previous season’s final stats was used to produce a system that correlates very well to 5x5 roto-scoring. That’s all well and good but it’s still essential to come up with a draft list incorporating principles intrinsic to points scoring.
If you play fantasy football, you know where this is going. The key to points leagues is rankings should not be based on raw points but rather adjusted points using the last player drafted at each position as a baseline. The idea is everyone in the league is credited with the number of points scored by the worst active player at each position so the person with that player essentially earns no useful points from that player.
Mathematically, find the worst draft-worthy player at each position, subtract those points from everyone at the position and re-rank according to those adjusted points.
Truth be told, this is by no means perfect, especially in a best-ball format. The calculation only works if one player occupies each roster spot all season – which is obviously not the case. In addition, the use of corner, middle, utility and players that are eligible for multiple positions skew the replacement level. Still, doing the best you can to determine replacement is better than ignoring it. Ultimately, draft flow comes down to varying expectations of player performance but having a starting point where, at minimum, the players are ranked accurately relative to each other is very beneficial.
HOW MASTERSBALL GENERATES CUTLINE RANKINGS
Let’s start with the easy part – pitching. There are ten teams with nine roster spots, so the expected points from the 90th highest total is subtracted from all the hurlers.
Hitting is where it gets dicey. Here’s what we know.
Players with multiple eligibility are assigned a primary position according to this hierarchy:
C > SS > 2B > 3B > OF > 1B
This is how I view the strength of positions – you may see it differently. Your team, your call.
The projected points for all the hitters are calculated. The top-140 (ten teams, 14 roster spots) are examined to see if the above criteria are satisfied, starting with catcher and moving the hierarchy. If a position is short, the highest ranking player at that position is brought into the top-140, knocking out the lowest ranked player at a position that has not yet been checked. When finished, the top-140 should now consist of ample players at each position to fill all ten active rosters.
The lowest ranked player at each position is identified and those points are subtracted from every player with that same primary position. These adjusted points are used to rank hitters and pitchers together.
To reiterate, this process isn’t perfect, but it’s better than using unadjusted points. Because of the unique Cutline points system, the adjustment isn’t all that steep. However, to those playing in points leagues other than the NFBC cutline, omitting the adjustment is the biggest mistake made. The projected points for hitters and pitchers will be computed and it is wrongfully concluded that one is way more valuable than the other based on raw points.
GO BIG OR GO HOME
Before we go on to discuss some specific strategies, it’s necessary to set the proper mindset. Sure, there’s a league prize, as the top-scorer in each ten-team league will pocket a nifty $250. Hopefully, it’s obvious that the NFBC Cutline is a contest where you’re....
Sorry friends, I need to save the rest for the Platinum subscribers. The remainder of this discussion along with Cutline rankings are now posted for Platinum. Please note we are very close to unveiling a promotion where you can get $20 off the regular cost of $34.95. If you qualify for the promotion, we'll gladly refund $20 off the purchase price once we are able to go live. To subscribe, click HERE. Please make sure you're registered and logged into the site. You can create an account via the log in box on the home page.
As you’re probably aware by now, the Mastersball crew spent last weekend in the Big Apple, participating in the Tout Wars festivities. Along with Brian Walton, I represent the site in the National League Auction. Instead of giving you the standard pick-by-pick review, I thought I’d offer my reasoning for a couple of purchases that I don’t usually make. This will serve as an excellent example of how really understanding your rules and the repercussions can aid on draft day.
By means of reminder, Tout Wars has a rule whereby teams are penalized a FAAB unit for each point they fall below a set threshold. For the first time since the implementation of the rule, I’m beginning the season with less than the maximum 100. I ended the 2014 season with 53.5 points which translates to 93 FAAB units.
Even though trading of FAAB is allowed, and we are able to get FAAB rebated for players put on the disabled list, I am going in with the mindset that I’m out of the running for having the hammer at the July 31 trading deadline, so I may as well spend early and often. Further, if I’m going to spend early and often, I should design my roster to best take advantage of this.
The two paths I opted to use were dedicating a couple of roster spots to $1 end-gamers as well as buying a couple of promising minor leaguers that aren’t likely to break camp but that should be up my mid-season. In essence, the prospects would serve as my trade deadline acquisitions, hopefully a bit earlier. Initially I’d use reserves to backfill the open spots, then I’d be aggressive with FAAB looking for an upgrade.
The $1 players I picked up are Charlie Culberson and Chris Denorfia. In reserves, I backed them up with Daniel Descalso and Brandon Barnes. I’m not thrilled that I’m locked into a shortstop only for Culberson having blocked the middle via buying a pair of second basemen in Kolten Wong and Daniel Murphy. That said, Descalso also qualifies at shortstop so I have the market cornered on potential Troy Tulowitzki replacements.
The players sent to the farm on my active roster are Jose Peraza, Maikel Franco and Robbie Erlin. There’s a very good chance that all three will be up and contribute this season. Peraza cost me three bucks, Franco was double that while Erlin was a buck as my last pick. Peraza is presently at swingman, so Barnes or Descalso could be promoted, or if I wanted to drop someone, I could FAAB a pitcher into the spot. Franco is occupying corner but will be replaced by Tommy Medica. I have better opportunity to upgrade Medica as compared to shortstop since either a first baseman or third baseman will do the trick. Assuming he’s ready by opening day, I took a shot that Joel Peralta would be selected to fill in for Kenley Jansen and get me some saves from a reserve pick, activating him for Erlin.
I really didn’t do a very good job of acquiring multi-position players, which would have aided this tactic. Other than Culberson and Descalso, the only two-position player is Scott Van Slyke, who is my sole first base eligible option. I’m fine with Van Slyke, but am obviously looking to upgrade the roster spot shared by my versatile Rockies.
I have no idea if this will be effective. Obviously, if Peraza, Franco and Erlin contribute, I’ll appear to be a genius. If not, not so much. But part of the fun of this is thinking outside the box and trying to get an edge above and beyond knowing baseball. It seemed apparent that figuring out a way to mitigate less FAAB would help me in the big picture. The scheme discussed above interweaves that with upgrading end-gamers and stashing a couple of prospects, something I usually don’t do. But the way I see it, the entire available player list is the equivalent of an extra roster spot for me since I’ll be liberally dipping into the free agent inventory. Assisting this is there are separate disabled lists, so I’ll always have reserves devoid of injured players. I’ll be short a reserve spot or two until my guys are hopefully promoted, but the player pool is not like a mixed league, there isn’t as much activity between active and reserves.
The entire Tout Wars results are HERE. I’m happy to address any specific player questions or elucidate strategy in the comments or forum.