EDITOR’S NOTE: Each point in the graph above represents a play in the game. Each point has a Game-Winning Chance (GWC) illustrated by percentage on the left side of the graph. Critical calls throughout the game are represented by green and red dots. As the game progresses, and time runs out, the team that started with a 66 percent GWC has achieved 100 percent GWC with the help of proper decision-making.
Probably the best metric available to coaches (data-driven or not) is a metric called Game-Winning Chance (GWC). It’s a real-time measure of how likely a given team is to win the game they’re playing. GWC is derived from countless variables, but according to EdjSports co-founder Frank Frigo, it’s a metric that every coach intuitively understands.
“Inevitably, at the end of a given game, one coach has no chance to win and the other coach has a 100 percent chance of winning,” says Frigo. “All along the way, there are all kinds of shifts where GWC fluctuates. It fluctuates because of decisions coaches make and player execution.”
A good analogy that many coaches are familiar with is a poker game. Think of a 1 percent chance of winning as a single poker chip. If one player gains a poker chip because of a good bet, then that player has a marginally higher chance of winning. At the same time, the other player loses a poker chip. It’s a zero-sum game. At the end of the game, one player has all the chips and the other player has none. This is just like a football game.
“The thing we realized is when we focus on GWC, it creates some really great insight,” says Frigo. “Traditionally teams think in terms of, ‘Oh, I have to score points. I have to advance the ball. I have to get a first down. I have to keep my opponent from doing certain things.’ And those all matter, but they matter in context. If you can represent these different facets of the game in a win probability metric, it gives you a much better understanding of how to manage risk, how to make good decisions and ultimately how to win.”
In a contest between two opposing teams, GWC becomes the commodity those coaches trade. Each coach is trying to get all of that commodity, while at the same time, reduce the risk that they will lose some of that commodity. It becomes a risk proposition.
“When a coach encounters a fourth-down decision, he’s in a ‘unique game state,’” says Frigo. “He’s at 4th-and-1 at midfield, it’s a tie score, there’s 8 minutes remaining in the third quarter. That defines the game’s state. There’s a GWC associated with that game state. What a coach is trying to do from there, if the model says he’s at 50 percent, he’s trying to move in the direction of 100 percent. He’s not going to get it all at once, but he can fractionally move in that direction with the right decision.”
RELATED ARTICLE: Football Analytics: Know Yourself, Know Your Opponent
GWC and the EdjSports analytics platform help coaches understand what types of decisions can help a coach steer his team toward 100 percent on a decision-by-decision basis.
Anthony Jones, director of operations for EdjSports, says that understanding what the analytics model behind GWC does can be boiled down to something very simple with real-world applications.
“All the computational power behind our platform just cannot be reproduced by a human, and that’s ok, because we know our model works,” he says. “It’s been vetted and tested very thoroughly by NFL and NCAA teams. It helps coaches know that, ‘Look, I don’t have to spend any brain power on this, for instance, if I’m deciding between a punt and go.’ It very quickly tells a coach what the relative strength of each recommendation is so he can understand what’s at stake, make a decision and get back to coaching.”
For more information, or to dig deeper into the advantages provided by utilizing GWC and other metrics, visit www.edjvarsity.com.