Chris Jones' third-down gamble countered probability, despite situational merit

Edmonton Eskimos' head coach Chris Jones is known for deploying unconventional strategies, and those have often paid off for him. However, a bold call Jones made late in last week's game against B.C. to go for a first down on third and four rather than have Grant Shaw attempt a 48-yard game-tying field goal didn't pay off, and that's spurred plenty of discussion this week. With the Eskimos taking on Montreal Thursday (7:30 p.m. Eastern, TSN/ESPN2), Jones' in-game calls will again be up for debate. So, was he right or wrong to go for the first down against the Lions, and what should he do in those situations going forward?

First, let's consider the general case of a third-and-four versus a field goal, with score and game time ignored for the moment. Current Toronto Maple Leafs' statistical analyst Rob Pettapiece ran a fascinating CFL study on all of the third-down data from 2008-2012 back in 2013, building on his previous work on the subject (which only considered the 2009 season). That study produced the following chart of the optimal decision from a probability standpoint at each yardage point:

As the chart indicates, going for it on third-and-four is generally only a clear best choice when inside the 10-yard line. It's a toss-up between going for it and kicking the field goal from the 10 to slightly past the 20, and then the field goal usually becomes the better decision.

How hard is it to pick up a third-and-four? The following chart, also created by Pettapiece based on that 2008-12 data, indicates the percentage at which teams succeed on converting any specific distance:

For longer situations such as third-and-four, there aren't enough data points from gambles to be useful (as teams only went for it five per cent of the time). If we presume they're converted at about the same rate as a second-and-four (an assumption Pettapiece wrote this week "pretty much holds in other situations"), we get third-and-four situations being converted 49 per cent of the time.

By contrast, how hard is the 48-yard field goal, especially in a dome environment like B.C. Place? TSN's Derek Taylor had some interesting tweets on that after the game last Thursday:

It's also worth noting that Eskimos' kicker Grant Shaw certainly has the leg for that distance, and that he's been perfect on the year on field goals (although he has missed some converts). A field goal that long is far from a sure thing, as Taylor mentions, but giving Edmonton a 67 per cent chance or better of hitting the FG if they attempt it seems reasonable. So, ignoring the specific situation, a 48-yard-field goal generally would seem to be a stronger option there than a third-down gamble.

In Jones' defence, though, there were situational merits to the third-down gamble. A field goal would only tie the game, and there was about a minute left on the clock. Yes, the Edmonton defence is impressive, but that's more than enough time for B.C. to string together some plays, and the Lions have a kicker/punter (Richie Leone) with his own massive leg. They would only need a single point to win, too, so that even puts the possibility of a game-winning rouge punt or a pass-and-punt in play. A successful third-down gamble would have had benefits for the Eskimos, too; it would have allowed them to keep their drive alive and potentially score a game-winning touchdown, and at the worst, it would have let them try the field goal from closer range with less time left on the clock. Thus, Jones' call is somewhat defensible, and it's easy to understand why he made it.

Overall, though, Jones probably got this one wrong. A 49 per cent chance of converting on third and four, with a failure likely meaning a loss thanks to the limited time left, is an awfully big gamble to take. Kicking the field goal wouldn't have been a sure thing, but it comes in at higher odds (likely 67 per cent or so), and it offers an opportunity to extend the game. Oddly enough, Jones has done this the wrong way around before, opting to kick on third and inches late instead of going for it. His gamble here wasn't all bad, but it didn't provide his team the best possible odds of victory.