The Fine Art of Tempo
PigSkinRevolution.com
The types of critical play decisions that ZEUS typically evaluates are transparent. This means when the offense lines
up to kick a field goal there is a high degree of certainty that this will be their play choice. The same holds true for kick-
offs, punts, extra-points and short yardage 4th down attempts. While exceptions (fakes) may occur, the offense
generally is deliberate and obvious in their actions and the defense is prepared. Because these types of critical play
choices are essentially binary, they can be assessed on their relative merits. Occasionally, critical choices may
consist of 3 legitimate options (i.e. punt, field goal, go for it) but this is rare. ZEUS is particularly good at weighing
such options and revealing the choice with the greatest return in GWC.
1st, 2nd and 3rd downs are not very transparent. The coach may have dozens of choices in his playbook to choose
from depending on the situation. While each coach in the NFL may employ their own unique strategies in any given
decision, their choices are tempered by the urgency of the situation. For instance, it is common wisdom to be more
aggressive when the clock is decaying and a team is trailing. Conversely, conservative approaches prevail when
leading. We refer to this adjustment in play calling as tempo management. Because ZEUS can evaluate the GWC of
any situation in a game, the model can provide great insights into tempo.
Assume we have 2 evenly matched teams facing each other. At the opening kickoff, each team will have a GWC of
approximately 50% and will be striving to reach a GWC of 100%. From this point forward, the GWC becomes a
tug-of-war. Every incremental coaching decision, play execution and random element will either raise or lower the
GWC. Here is an example of what we call an equity progression chart from a previous AFC championship game.
ZEUS has assessed each teams winning chances at each play increment during the game. For instance at play 115,
the leader (New England) had a GWC of about 80%. With about 45 plays remaining in the game, they would need to
add about 0.5% GWC on average during each each incremental play for a linear progression to victory. Their
opponent on the other hand would need to average 4 times that amount or about 2.0% GWC each play. In reality, we
know these types of progressions do not occur in a linear fashion.
Typically, play choices with greater rewards carry greater risks. A 'hail mary' pass may be great when it works but it is
almost certainly not the best choice with a tie score early in the 1st quarter. Trailing by 6 points with 1 second
remaining in the game at mid-field - and it makes a whole lot more sense. Balancing aggressive and conservative
actions with the inherent risks is the art of tempo management. A linear path to victory serves as a general guide to
making good tempo choices. In reality, it may be correct to attempt to grab more or less GWC than what the linear
guidelines would indicate. On 1st - 3rd downs, ZEUS can evaluate the effectiveness of various general offensive play
distributions including 'take a knee', run, short pass, medium pass and long pass. Each of these choices has its own
unique distribution of outcomes and risks which can be adjusted for the caliber of the offense and the opposing
defense. While ZEUS may not suggest which specific "medium pass" play is correct, it can still indicate that this
category of play is optimal based upon tempo considerations.
Lets look at some examples from monday night's contest between the Patriots and Vikings:
At the beginning of the 4th quarter the Vikings are in the nearly hopeless position of trailing the Patriots 31-7. They
have the ball on their own 34 yard line with 15 minutes remaining in the game. On average, approximately 35 plays
will be executed before the end of the game. The Vikings current GWC is a dismal 0.4%. Therefore they will need to
make up 99.6% GWC over the next 35 plays to pull this out. On average, that is a whopping 2.84% GWC per play.
So for a linear progression to victory, they must raise their GWC from 0.4% to about 3.2% on the next play. Lets
assume that 25 seconds transpire and let ZEUS determine how many yards the Vikings must gain to set forth on a
reasonable path to victory. It turns out that even if they gain 65 yards on this play (now 4th and goal on the 1) they can
only get to 1.7% GWC. In fact, a touchdown with a 2 point conversion still falls short of the linear figure at 2.7% GWC.
Does this mean that it is mathematically impossible for the Vikings to win? Of course not. A football game is a
volatile endeavor and progressions are almost never linear. Additionally, each play whether on offense or defense,
does not always provide enough equity to progress toward victory on a linear path. Some plays provide greater
opportunities and some, as in the case just described, provide less. But it does illustrate just how big of a hole the
Vikings had dug themselves. Any thing less than a crack at the end zone right now is sub optimal. The risk
parameters have reached a point where they should choose whichever play in their book has the highest probability of
immediately scoring a touchdown regardless of risk.
Now lets look at things from the Patriots perspective a little earlier in the game.
With 9:45 remaining in the 2nd quarter, the Patriots began their drive at their own 29 yard line leading 10 - 0. Their
GWC is approximately 78% at this juncture of the game. This means they must grab an additional 22% GWC to
secure victory over the next 40 minutes or roughly 100 plays. A simple calculation reveals that the Patriots, on
average, need about 0.2% incremental GWC to move toward victory on a direct path. How does that translate to
yardage on their next play from scrimmage? Again we will assume that about 25 seconds will tick off the clock on
average during the next play. This means the Patriots need to gain 4 - 5 yards to move linearly toward their goal of
100% GWC. We are not suggesting that a simple running play is the clear choice here. However, the tempo analysis
suggest that unless the Patriots have a compelling reason to do something more dramatic, they should choose a play
with the least risk and highest expectation of gaining a minimum of 4 - 5 yards. In the actual game, they completed a
12 yard pass for a first down which advanced their GWC by 1.6% whereby an incomplete pass would have cost them
0.6%. According to ZEUS, a medium pass play was in order in this situation and the Patriots not only made the
correct choice but executed successfully.
The day is coming, in fact we think it may have already arrived, when a computer model such as ZEUS will be able to
call each play of the game better than any human. No, we are not suggesting that the coach is becoming obsolete.
Someone needs to script the plays, coach the players in proper execution, and the quarterback will always need to
have the ability to read a defense and call an audible. The construction of the play-book 'menu' will likely always
require a human touch. The optimal 'ordering' is quite another matter.