Within the decisions that are made in this great game of football, there have been many theories on how to go about making the right decisions in critical situations. While coaches learn from their mistakes and get lucky a lot of the time, past history is not enough to rely on when making future decisions.

When speaking with several of my Quantitative Analyst friends, I call them number counters, theory thinkers and calculator crunchers. I have dabbled a little in the MoneyBall comparison of statistical numbers vs. scouting performance myself, working to make a good decision when drafting. As explained to me, there are two kinds of analysis when it comes to making tough critical decisions. First, Descriptive analysis is what coaches actually do. Second, Prescriptive analysis is what people should do. Rarely are the two things the same.

When coaches self-scout themselves on their performance and production, they are looking for what is called a win probability model. When evaluating third-down decisions vs. distance (example: third-and-5, third-and-7), on both sides of the ball (offense and defense), this is called prescriptive analysis. When working to explain what coaches are actually pondering when making a decision is called descriptive analysis.

The computer program crunches the data entered and spits out the results called a matrix, which are studied and restudied to find tendencies that their opponents are looking for as well. When coaches are in the heat of battle, competing in a game at a high level, quick and precise decisions need to be made in a matter of seconds. Most if not all collegiate and professional coaching staffs up in the coaches box have data charts ready for all scenarios. Some, if not all, teams use a computer or the latest device used is the IPad2 to display instant data.

Professional teams actual have off-the-shelf software or create in-house programs to take away and dilute the human error in certain situations. Let’s take a look at some situations and ways and reasons why coaches make certain decisions, and on how the prescriptive analysis can make a difference in live game action.

Self-scouting breaks down a team’s situational history. While there are no brain surgeons in the NFL, the game has developed into a cat and mouse conservative chess match at an intellectual level. The majority follows past wisdom from previous coaches, mentors and peers, keeping structure of coaching and decision-making close to the vest and remain with the conventional thinking and rarely deviate from their upbringing.

When cooking or barbecuing, this approach makes total sense, but the world of the NFL is far less predictable then coaches and owners for that matter would like. Coaches like to settle down and make decisions based on past outcomes and favorable decisions, rather than lead their teams into uncharted territory.

This is where the self-scout matrix comes into play. Based off field position, down and distance and production there is probability, along with the possible consequences ending in poor results. Keep in mind that several coaches will take the chance to create production in awkward decision-making outside the norm in order to increase performance and generate production because for a split second they forget or ignore the certainty available to them.

One of the more conservative approaches or decision-making methods I call “best-play ” mode. It is based off history and assures you some production, such as a coach’s script plays in order, based off down, distance and formation.

Another decision method is known as the “play-it-safe” mode. This method of thinking seeks to minimize potential mistakes. This is where play-calling is extremely important based off the self-scouting matrix, but also the charting of data in the present moment, that may show the defense for example doing something different then what they have shown in the past few games. If an offensive coach has plays set for a defensive situation, what are the odds hypothetically of gaining production. The computer should kick out a data of possibilities and percentages.

Know this is when human error comes into play. Coaches at all levels of competition rarely pay attention to the matrix of possibility; they pay attention to what they practiced all week, what was taught to the players Wednesday through Friday, and whatever happens coaches brains shut off the data in front of them and make poor decisions based off emotions that plays a major role.

The self-scout matrix helps reduce the uncertainty in which the defensive front and coverage the offense will face based of formation, personnel and motion. Even though the offensive coordinator can’t predict exactly which one they will see, they can estimate the probabilities of what to expect. The expected defense is a weighted average of the possibilities. Again, the matrix data tells the coach they have a one-in-three chance of facing a certain defense based off the formation they align; now they can estimate the expected defense. For each play call, you look at the new defenses that faced a certain call, which expected front and coverage based off pervious series resulted in positive yardage from Play X is (1/3)(-4) + (1/3)(4) + (1/3)(12) = 4. The expected production for Play Y is 3, and for Play Z is 2. The expected production by play elimination says Play X is the best choice.

These methods each call for a different decision. Each method is logical and consistent based off the self-scout matrix, where there is truly a correct method that counts in this great game of football only one prescriptive analysis. With the self-scout matrix, coaches know the probabilities and the possible production, or feel good about the current production results.

The challenge in this game is knowing the possibility of production through yards, and points. Keep in mind this theory only helps if you have the lead, because if it is third-and-6 and you’re losing by four points with the clock less than two minutes in the fourth quarter, a field goal doesn’t help you.

This is where the win probability is huge. Like all sports, winning is the bottom line, nothing else matters! As my friends and I discuss, win probability is perfectly linear, which is expected through production analysis. Where thought patterns are miscued is in losing. The fear of a loss makes some panic and human nature sends everything out of whack.

When you look at the last five Super Bowl teams, you will find those that were winners, had the advantage in decision-making, based off expected production from the self-scout matrix and currently charted live data, made sound decisions resulting in production.