One thing it took me a long time to learn is just how to compare
the relative value of different positions.
Do not make the mistake of thinking that the point totals at a position
being the only relevant comparison point.
If that was true, then everyone playing in leagues with standard scoring
would draft kickers in the first round.
Instead, the range of points should be based on the number of points
from the top player at the position down to the replacement level player at
that position. For kickers in a 12-team
league, where a team would only get a second kicker to cover a bye week, the
range of points at that position may only be 10 points. So even though the best kicker may score 150
points over an entire season, no one would make any type of significant investment
in acquiring that best kicker when there are better relative values at other
positions.
The value of a player to a team is the difference between his
points scored and the points that could be obtained by replacing that player
with a replacement level player. I
initially thought only starting lineup requirements defined player value. If I was in a 12-team league that required 2,
and only 2, starting RBs, then I cared primarily about the relative value of
those top 24 running backs. Maybe this
would produce optimal results if no players ever became hurt and we were
perfect at predicting who those top players would be each year, but you’re just
as likely to predict the winning lottery numbers. I then switched that valuation to roster
requirements. If teams had, on average,
7 WRs on the roster in a 12-team league, then I looked at the relative value of
those top 84 WRs. But the advent, and
expansion of, flex positions, along with my penchant for playing in very deep
roster leagues, led to some obvious mistakes in how I valued players in such
leagues using that format.
I have now settled on the replacement level player being the key
in player valuation. So how do we define
that term?
I define the replacement level player as the player the aggregate
team owners turn to when they make a change in the group of players they
consider using for their starting lineups.
At such a position as placekicker, the replacement level player is easy
to figure out. Almost no one is going to
roster a second kicker, so in a 12-team league, the 12th best kicker is
replacement level. Not 13th mind
you. If a team finds themselves without
one of the top 12 kickers on their roster, they cut him and pick up one of the
12. I.e., one presumes the replacement player
is superior to the previous starter, i.e. the 12th best, and thus that player is
the replacement level kicker. In leagues
with very limited roster spots, there are quality players available as free
agents and a team will often turn to one and insert them directly into their
starting lineup to cover injuries, bye weeks, or unexpected poor performance
from an already rostered player. As
such, the league’s roster requirements can easily be used to determine the
replacement level player. In a 12-team
league where only 18 tight ends total are on rosters, the TE18 is the
replacement level player.
In deep roster leagues, there may be no free agents worthy of
adding to one’s roster, especially at the offensive skill positions. Such teams may have a group of players already
on their rosters that they would not normally consider inserting into the
starting lineup, not even if one of their starters goes down injured. In such a situation, the team owner may
rather trade for a replacement from another team than use one of their own lesser
players. Determining that delineation
between useful and replacement level player in such deep roster leagues is
tricky. If one were to look at one’s own
league’s transactions and starting lineups from the prior season, one could
spend the time looking at which players were selected for starting lineups, how
often, and which were not, and determine an answer. If one is setting up a new league, then
you’re just guessing. I say that,
because the replacement level is determined by how the individual team owners
make decisions, and every team owner makes decisions differently. Nevertheless, our best guess is better than
the alternative, and is absolutely essential in setting up the best possible
rules.
Because I value players based on the difference between their
projected points and a replacement level player’s points, and because we want
the positions to be valued as close to equally as we can to maximize our
strategic options as built into the league rules, if we can manipulate which
player is replacement level, we can use that as a valuable tool in creating
that equal valuation and minimize the distortions from having differentiated
position scoring.
The first method in manipulating the replacement level line is
with starting lineup requirements. For the
clearest example, let’s look at the difference between Superflex and
non-Superflex 12-team leagues. In
non-Superflex leagues, teams likely have 2 quarterbacks on their roster, and
sometimes only 1. Teams with 2 QBs have
the extra to cover bye weeks or protect themselves from poor play or perhaps
alternate their starter based on matchups, whereas teams with only 1 think that
an injury replacement NFL backup may be just as likely to be better than an
existing free agent. So in such a league
the replacement level QB could be the QB20, i.e. 2/3 of the league’s teams have
2 QBs rostered whereas the remaining third have only 1 QB and will only add a
replacement level second QB to cover a bye or injury. If I project the QB1 to score 150 more points
over the coming season than the QB20, then I will value him as such in
comparison to other QBs and players at other positions. In a 12-team Superflex league, the demands in
having 2 starting QBs as well as (hopefully) another to cover 2 separate bye
weeks will cause demand for more QBs than there are starting QBs in the
NFL. Fantasy teams will likely roster
NFL backup QBs with a possibility of seeing playing time, or playing well if
the starter is injured. In such a
league, the replacement level QB is likely QB33. Going back to the same hypothetical
projections, the difference in anticipated points between QB1 and QB33 could be
450 points, or 3 times that from our non-Superflex example. By adding the second QB starter, we have pushed
the replacement level down 13 spots and tripled the relative value of
quarterbacks without doing a single alteration to our scoring system.
Similar manipulations may be made by increasing, or decreasing,
the starting position requirements at other positions, or adding flex positions. I am a proponent of limiting the number of
starting running backs but increasing the number of starting tight ends for
this very reason.
Another way to manipulate replacement levels, and one in which I
expect the most pushback, is putting a cap on the number of players one can
roster at certain positions. I’m a big
proponent of IDP leagues, and am rather picky about leagues I agree to
join. A couple years back I set up a
16-team dynasty league using very similar rules to TUFFSCCL, including starting
a full 11-man defense, but with one major difference. I decreased the roster size from 53 to
40. In my mind, I was thinking that this
would increase the difficulty and frequency of roster decisions, which I
believe is necessary for dynasty leagues that can otherwise remain too static
and inactive. In prepping for the
initial draft, I mistakenly kept the player valuations the same as in
TUFFSCCL. This was an enormous mistake
because teams placed higher value on backup offensive players than defensive
starters. Even after two years,
starter-worthy defensive players are readily available on waivers, whereas at
least a 3rd round draft pick, and often much more, is needed to trade for a low
scoring RB you’re interested in taking a flyer on. To rectify that problem, we are transitioning
to a larger roster with position limits.
Similarly, last year I agreed to take over an orphan team in an
established league based on, among other things, assurances that defensive
players were highly valued. The team I
inherited was obviously lacking in multiple areas and was in need of a rebuild,
ideally by trading away established players for multiple draft picks (since my
team had no rookie picks whatsoever).
The best player on my roster was S Landon Collins, the highest scoring
IDP from the previous year by a significant margin. He was also young, under a long-term
contract, and the Giants looked to continue using him in the exact way in in
which he had acquired so many points. So
I put him on the trade block with the hopes of acquiring multiple picks. In my TUFFSCCL league, he would probably be
valued as more than a first round rookie pick.
In this league, however, the best
offer I received, despite significant efforts to drum up interest, was for a
single 4th round draft pick, or less than what probably-should-be-retired RB
Jamaal Charles went for. Now I will
gladly admit that, based on the league’s rules, this was a mistake that the 15
other teams in this league were making, as I was able to improve my roster
focusing on these undervalued defensive players, picking several up in free
agency for minimal cost. And if other
teams realize that there’s a competitive advantage to be had there then the
relative values of IDPs will slowly improve.
But maybe not, because it proved just about impossible for me to acquire
any offensive players, and I had to
significantly overpay for mediocre talent just to put together the minimum
necessary to have a starting lineup, which hampered my competitiveness.
When designing your league rules, player valuation is set by “the
market,” i.e. the actions and behaviors of the league’s owners. In both of these leagues, higher valuation
was placed on crappy offensive players than good defensive players. If your intent is to make having a top
defense a legitimate path towards winning your league, then you should strongly
consider capping the number of RBs, WRs, TEs, and perhaps even QBs, that a team
can have on their roster. By doing so
you are artificially creating the replacement level delineation at those
positions by making free agents available for weekly pickups. At the same time, you increase demand for the
other positions because teams are going to fill up the rest of their
roster. And by doing so, you’re making
it easier for them to consider those players as a starter based on matchups /
byes / injuries, which will deepen the replacement level line at those
positions. So we increase the spread of
points between the top defenders and replacement level players, and lo and
behold, we have suddenly increased the relative value of those positions.
*** MFL.com has a league option at my request that can also be used to increase the number of valuable players. Because almost every player on an NFL team’s active roster contributes in some way on the field come gameday, I requested that a fraction of non-starting player’s points be added to the fantasy team score. For my league’s, I’ve set them at 25%. Also, to prevent teams adding additional players at specific high-scoring positions, such as QB and PK, my leagues use the “taxi squad” feature on the MFL.com website to differentiate between “active” and “inactive” players (just like an NFL team), and require ALL non-starting QBs and PKs to be on the inactive list. Since teams try to reserve those spots for injured players or players on their bye weeks, this is sufficient to remove any artificial demand for those positions. The set-up has worked very well. The rule rewards teams that actively churn their roster to score points from top to bottom. It especially increased the value of CBs, who in almost all IDP scoring systems are the least valued position due to the large number of them accruing very similar statistics each week. Since even an NFL team’s third “nickel” CB sees the field a lot in today’s pass happy game, that’s 96 players at that position get fantasy points each week. My leagues’ teams learned to load up on CBs to fill out their roster in order to maximize their bench points. Thus the replacement level CB is much deeper than at all other positions. This also increases the point range at the position, making players at the top end of that position more valuable, and comparable to other IDP positions.
So how do we put it all together, combine scoring rules, differentiated by position, with starting lineup requirements and roster requirements, to create your ideal, competitive league with multiple paths in crafting a competitive roster? However you want. No, that’s not a cop out. It’s your league. Make the decisions that fit best with your desires. If you want to check out mine as an example, check out my TUFFSCCL league at MFL.com.
Here is the relative value of the positions in my TUFFSCCL league. Already you can see one problem as 2016 was a particularly good year for RB scoring, especially at the top, compared to prior years. Since I started the league in 2014, and designed the scoring system primarily on fantasy statistics from 2011-13, last year's RB performances appear to be a significant deviation from the norm. Though 2017 Todd Gurley is doing his best to say the era of the RB has returned. Nevertheless, you can see that RB, QB, and WR are pretty close, with TE just below (though with a big gap at the top of the position). Then comes the IDP positions clumped together, and pretty close, though giving an edge to DT and DE at the top end as well. With 11 defensive starters compared to 1 QB and 7 offensive skill position players, it creates about even point totals between offense and defense. I also have the Team OL position ranked amongst the defensive players, making it fantasy relevant. PK and PN are part of the game, so I account for them. But I intentionally minimized their points scored, and you can see that in the chart.
Feel free to email me at earlyr@gmail.com or find me on Twitter at @RyanAEarly and I’ll be happy to discuss, offer suggestions, or critique your ideas. You’ll be doing me a favor as otherwise I may just take over another orphan team simply out of boredom.
No comments:
Post a Comment