Analyzing the impact of close games on a team’s record in the NFL

by Jason Pauley

Watching Baltimore fail on a couple of 2-point conversions at the end of the game and lose to both Pittsburgh and Green Bay in the last few weeks got me thinking about close games and their impact on a team’s record. What teams have had the most positive and negative impact from close games this year? Is winning a close game skill or is it luck? In other words, do good teams win close games more than bad teams, or is it a toss-up?

In this post, I’m going to look at the impact of close games on a team’s record through game 14 of this season. What teams were helped and hurt the most by their record in close games? It’s important to note that choosing 3-point games is somewhat arbitrary, but I chose that number because it’s essentially a difference of a field goal. Also, not all 3-points games are equal. A team may be down by 10 and score a last-second garbage-time touchdown making the game a 3 point game. But for the most part, it can be assumed that the vast majority of 3-point games were close matches that could have gone either way.

Let’s start by looking at the rate of close games for each team. This year, 23% of all games in the NFL have been decided by three points or less. But the Ravens…they are in a league of their own. With 50% of Ravens games being decided by 3 points or less, I have to wonder what this team has done to their fan base’s life expectancy this season (maybe that’s an off-season post).

Here are the % of close games for each team through 14 games:

The Lions, Vikings, and Bengals have also experienced a high percentage of close games. Between weeks 3–12, 50% of Detroit’s games were decided by 3 points or less and none of those games resulted in a win (4 losses and 1 tie). Although the Lions aren’t good, it shouldn’t be a surprise that they are beginning to put together some wins. They were overdue and have had their share of bad luck.

Speaking of luck…is it luck, or is winning close games a skill? One way to determine this could be to plot all of the teams’ records in games decided by 4+ points vs all of the teams’ games decided by 3 or less. Just looking at this year, the chart below shows each team’s winning percentage for close games on the Y-Axis, and for games decided by 4+ points on the X-Axis. My assumption here is that if winning close games is a skill, then a team’s winning % in close games would correlate somewhat with their winning percentage outside of close games. But what you see below is that there is virtually no correlation between the two.

This data begs for a larger sample size and I plan on going deeper on this in the off-season. I don’t expect the findings will be different, but it’s worth looking at more than one season’s worth of data. This chart implies that winning close games is more about luck. I would argue that maybe the best teams avoid close games altogether, but when they are in a close game, it’s a toss-up. Despite what this chart is showing, I’m not sold that it’s entirely luck. Given a larger sample size, we might see that some quarterbacks may have that clutch gene and will win a disproportionate amount of close games. Some coaches might choke more than others. Having an all-time great kicker might also help in close games. However, I do believe it’s more luck than skill.

So the Ravens have a lot of 3-point games, and recently those games haven’t gone well, but over the entire season, they are 4–3 in those games, so their impact has been fairly even. Poor Seattle though, they are 5–9 this season, but in close games, they have a net win differential of -4. They are 5–5 excluding 3-point games and they are 0–4 in 3 point games (two of those losses in OT). Excluding 3-point games, Seattle is a team in the playoff hunt, including 3-point games, they have the 4th worst record in the NFC. Pittsburgh on the other hand is a team in the playoff hunt at 7–6–1. However, they have had a net win differential of +3, with a 3–0–1 record in 3 point games. Their .400 winning percentage excluding 3-point games would have them eliminated from reasonable playoff chance at this point in the season, but here they are one win out of the 7th seed.

Here are 3-point game records for each team grouped by win differential from close games.

The teams at the top of the list have had the highest net wins from 3-point games

The table above just shows each team’s record in close games, but it doesn’t show how much each team was helped and hurt by their close game performance in terms of their winning percentage. A lot depends on their overall record and what percentage of their games were close games. Just because a team is towards the top or bottom on that table above, that doesn’t mean they will definitely be towards the top or bottom in the chart below (although in many cases the two tables will correlate). An example of this is Jacksonville. They are towards the middle with a +1 win differential on the table above, but the impact from close games on their overall record is the 4th largest in the league. This is because they have won only two games this year, but both of their wins were within 3 points; the 23–20 win against the 1–5 Dolphins in London, and then a shocking 9–6 home win against the Bills.

The Falcons have enjoyed the biggest positive impact from close games as a result of their 3–0 record on 3-point games. 50% of their wins have come from 3-point games, which is twice the league average. At the bottom of the list are the Bengals who have seen 67% of their losses come from 3-point games; almost 3 times the league average. Cincinnati is still sitting at the top of their division, but they have a .750 winning percentage excluding 3-point games, and a .571 including 3-point games.

Below are what the standings would look like if we were to exclude close games (3-pts or less) from the data. Given that 23% of games are decided by 3 or fewer points, the impact on overall standings isn’t huge. But the luck that teams face in close games can certainly move them up or down a few playoff seeds, move a borderline team into or out of the playoffs, or have draft position consequences. I’m using win % instead of record in the hypothetical standings because teams have played a different number of close games.

Hypothetical standings through game 14 if we exclude games decided by 3 points or less

I believe that close games are a combination of luck and maybe some skill, but mostly luck. Segmenting wins this way can be useful in projecting within-season regression or perhaps year-to-year regression. It’s also fun for fans to have something to be thankful for or to complain about. When Bill Parcells said “you are what your record says you are”, well that’s true, but it doesn’t mean you will continue to be what your record says you are. A team’s record in close games can be an indicator of which teams are stronger or weaker than their record, and might suggest which teams may have some positive or negative regression in the near future.

Additional info including some of the data in charts and tables above + some new information