Welcome to Part 3 on my series on value based drafting.
I have gotten a lot of requests to get the articles out before people
have their drafts, so I am combining the last two articles into
this one. I will also try and keep less to the technical and more
to the ideas, so you can get this and apply it right away.
I had a couple of questions on my last articles about how to approach
the modifiers I gave you. For example, if wide receivers 1-12 have
a modifier of 70 % and wide receivers 13-24 have a modifier of 80%,
does that mean that receivers 13, 14, 15 are worth more than 10,
11, 12? Of course not, but I wasn't clear about how to deal with
this problem. What I recommend is looking the numbers as one smooth
curve, so if the average of one is 70%, and the other 80%, then
where they meet it should be about 75%. Remember to take this stuff
with a grain of salt and use the idea's behind it, not just the
numbers I give you.
In today's article, I will be going over different techniques that
I call "flex" value based drafting. These are ways to adjust VBD
in order to adapt for different strategies and other outside aspects
of the draft. I call it "flex" drafting for two reasons - it is
a stronger system for drafting than normal VBD and it is also more
adaptable, or "flexible." We will make three types of adjustments,
for player utilization strategies, for other owners, and in draft
recalculations.
Player Utilization Strategies
One approach for roster management is to draft several players to
fill a single starting spot. For example, if I don't draft a top
shelf quarterback, I might wait and draft three quarterbacks in
the middle rounds. Then, depending on potential breakouts, hot streaks,
and quality of opponents, I will rotate all of the quarterbacks
during the season. That way, I can get top quarterback production
by expending three mid-round draft picks instead of one high pick.
There are a couple of pitfalls with this strategy, the biggest of
which is the question of how accurate can you really be in predicting
game to game performance? Just because Kerry Collins is playing
the Minnesota Vikings, does that really mean he will produce more
than Charlie Batch in his games against the Tampa Bay Buccaneers
dominating defense?
The seemingly obvious answer is yes, because over a 16 game schedule,
Minnesota will let in about 80% more yards than Tampa Bay will.
But as I've mentioned before, neither real nor fantasy games are
based on a 16 games totals - they are played game to game. So I
took a look at the 128 regular season games played by eight worst
pass defending teams. The question I asked was how often did a quarterback
achieve more than his season average in those games? The answer
was that in over 85% of the games, the quarterback achieved higher
results against the bad defense than there season average. The cumulative
total was an increase of 29% for all quarterbacks, 36 % for those
quarterbacks drafted from the fifth round to the 14th round in last
years drafts.
So that is a useful and reliable roster rotating technique; however,
is there a way that it applies to VBD? In short, yes. That is because
there are some positions which you cannot wait and expect the same
results. For example, in most leagues, there will not be enough
running backs left, after the fifth round, to be able rotate them.
So it become more important to get a top running back since there
is not an alternative if you don't. The same holds true for kickers
and tight ends, but for a different reason - they are almost as
likely to score less points, as they are to score most points. In
other words, at least under the criteria I am using, they are not
predictable. Every other position was.
Simply using the criteria of the ranking of the opponent, I found
I could increase the ratings of quarterbacks (as mentioned above),
wide receivers (37 %, and 39% percent for receivers outside the
top ten), and defenses (nearly 60% fluctuation). Relative to these
positions, rotating where you have to get a top guy is that much
more important. So how do you put this in VBD?
Well, if top RB, TE, and K are more important (because of the lack
of alternatives) than QB, WR, and D's than you can either say that
the first three positions are worth more, or, that the last three
positions are worth less. I think it is easier to calculate the
latter of those two, so each position would be worth that much less
than they can be increased through rotating. So quarterbacks would
be worth the inverse of their 29 increase, which would be 100/129
or a 22 % decrease. For receivers it would be 100/137 or a 27% decrease,
and for defenses it would be 100/157 or 37%.
Now, there are some bad traits about this. For example, for the
85% that met or exceeded there season average, 15% didn't. Additionally,
drafting three quarterbacks in the middle rounds takes three roster
spots that could have been used for other quality players, where
as a top quarterback and his backup only take up two spots. So,
going with my theme of taking things with a grain of salt, I only
modify player's number by about half of the decreases I mentioned
above. But that decision is ultimately up to you.
Average Draft Position
Another consideration that I think is really important is where
will a player get drafted if you don't draft him. For example, tight
end Tony Gonzalez is my 11th highest rated player, but I know I
can almost always get closer to 25th overall, so I don't need to
pick him that early. Furthermore, some of his counterparts, Wesley
Walls, Rickey Dudley, and Shannon Sharpe have been dropping into,
and beyond, the middle rounds. So despite how highly I consider
those players, I pick them one or more rounds later than their rating
dictates their selection - just so long as a can get them. Most
major sites have draft trackers, which will tell you the average
draft position that each player gets drafted. You should look these
over for players that you like a lot but that might not need to
be picked quite as highly.
There isn't really a formula for this, but I never will pass up
a player if the player that I would draft instead is rated more
than 20% lower on my ratings. So if I rated Gonzalez as #11, I will
pick the 12th and 13th rated player over him if I think I can get
Gonzalez later, but not really anyone much deeper. But that is just
my own "risk" analysis, you should figure out your own.
Another thing you should consider when you are making this type
of decision is what type of draft you are participating in. In order
to develop my own skills, I participate in an inordinate number
of leagues, so I decided to try and summarize some of my experiences
in different levels of experiences.
Beginner Leagues: The big characteristic of beginner league's
is that just as many quarterbacks are taken early as there are running
backs taken. Wide receivers also creep up a lot, particularly you
will find receivers taken in the second though fourth rounds that
would have gone several rounds later. When drafting in beginner
leagues, you can usually wait longer for running backs, and to a
lesser degree, tight ends. Also, the players will be in just about
the same order as they finished the last season. But then again,
if you've read this far in this series, you are going to clean up
in any beginner leagues your in.
Veteran Leagues: In veteran leagues, besides better picking,
wide receivers drop and tight ends seem to rise. A veteran level
league also will deplete itself running backs totally within four
rounds, and often times within three. In a veteran league I will
try and grab three good running backs early because they aren't
going to be around later. While the top tight end will go right
earlier, one of the top four seems to still last just as long as
in beginner leagues.
Expert Leagues: Now these are the fun ones. I am defining
expert as any "traditional" expert league on the internet, or any
league made up of really committed veteran owners. I wouldn't expect
to get a decent running back past the first two rounds, but sometimes
in the frenzied running back feeding, some good values at other
positions drop. If you think you have an ace in the whole who you
can get later at running back, don't be surprised to see your top
rated wide receiver or top three quarterback last all the way to
the end of the second round. Also, we "experts" seem to take the
lesson of not taking a kicker or tight end early way to seriously,
and sometimes you can grab some really good picks there. Don't expect
to get a top tight end outside of the fourth round.
In-Draft Recalculations
In one draft I was in this year, I grabbed Tony Gonzalez in the
third round. Satisfied, I continued drafting as normal until it
got the seventh round and Wesley Walls was still on the draft board.
Since he is in the mid-twenties on my draft board I should have
drafted him ages ago, but I hadn't because of Gonzalez. Not wanting
anyone else in the league to get that good of a deal, I took him
as a backup. Should I have?
He was the highest rated player available by about 25 spots, but
he will probably produce little to zero points backing up the explosive
Gonzalez. Furthermore, since he dropped that far in the draft, I
have serious questions about what his trade value would be in that
league. This is the type of problem that we don't really have a
solution for, particularly in a live draft. However, if you participate
in a draft that even has ten minutes between picks, there is a solution.
Another problem, which conveniently has the same solution, is how
to adjust given runs on different positions, particularly when the
players drafted aren't who you would have drafted, leaving a player
on an value "island" for his particular position.
The solution for both of those problems is figuring out how to re-figure
player value's during a draft. For example, I set up my baseline
to draft three starter quality running backs. If I pick two running
backs in the first two rounds, that means that I only want one more.
So instead of continuing to draft with the numbers as they are,
I can re-do my values with a baseline as if I only wanted one more
starter. What I would do is eliminate all of the players who have
been drafted, and re-do the baseline for one starter out of who
is left. In a league where everyone else is getting running backs
as fast as you are, this shouldn't effect the ratings too much.
But if there are still a lot of quality backs that are worth about
the same, the VBD will tell you (accurately) to wait.
When you do this, you should eliminate all players who have been
selected, and then adjust your baselines for each player you have
selected. So if ten receivers have been taken, but you haven't taken
one, the baseline would be still be the same number down, but a
different player, probably ten spots lower. This of course will
raise the value of each receiver, which is exactly what you want
to happen.
In the example I mentioned about Wesley Walls, if I started with
a baseline of 18 players, I would have knocked it down to the 6th
best remaining (minus the 12 for one starter gone, leaving me an
expectation of maybe drafting a backup). Truth be told, I think
Walls was still worth it, but if I re-figured the results I would
have known for sure.
Practically, you can only use this technique if you are drafting
in a draft that stops every round, or is an elongated email draft.
But, if you are, it will improve the accuracy of your VBD incredibly.
No other draft system has come close to being able to adjust that
accurately during a draft.
If you won't have time to make re-adjustments in the draft, you
can make before to help with certain situations. For example, the
after the first three rounds I pretty much know I will have two
running backs and a tight end- I just have just gone through enough
drafts to see it. I also know that about seven quarterbacks, twelve
running backs, ten receivers, and one tight end will have been taken.
So I can re-adjust for that number of players missing and my new
baselines to take into account my existing roster. That way, for
the nine times out of ten that situation happens, I have a new sheet
to refer to.
The Conclusion
What I hoped to demonstrate in this series of article is how you
can utilize value based drafting to take into account not only your
personal projections of players, but also factors like predictability,
strategy, draft position and changing situations. You can involve
just about any factor with value based drafting, you just need to
be able to find a way to put your beliefs into numbers.
I want to thank everyone for reading and for sending all the kind
emails. Good luck this year, unless you are in a league with me...
just kidding, you can be as lucky as you want and it won't help.
