I have a pretty unbelievable story to share regarding the 2017 Colts. This team has largely been forgotten, because they didn’t have Andrew Luck & the coaching staff was fired after the season. This is a story about the consequences of horrendous decision making.

This is what happens when a team doesn’t pay attention to detail. This is what happens when they don’t even know the details because they haven’t studied them. They haven’t studied them because they haven’t bothered to crunch the data and realize there are details to be found. That’s analytics. The use of math or statistics to gain valuable knowledge from data. Knowledge which can be used to recommend action or guide decision making. Literally, the Colts didn’t know what they didn’t know. But if they bothered to break it down, here is what they would have learned. And perhaps they would have learned it in time to do something about it. To win more games. To save jobs. To change their season:

The 2017 Colts finished as the 3rd worst team in the league, per the standings, and they won only 4 games all season. That much you know. You probably don’t know any of the rest. Hell, the Colts themselves probably don’t know most of it…

(NOTE: I shared this story on Twitter. Presenting the data in a step-wise manner (like tweets in a thread) seemed easier to digest rather than compacting in paragraph form, so I maintained that format with bullet points below)

  • The 2017 Colts may have finished 4-12, but in their 16-game season, the Colts trailed at halftime in only 6 games. They held halftime leads in 9 games. Yet they went 2-7 in games they led at halftime.
  • The 2017 Colts are the ONLY team in the last 27 years to lose at least 7 games which they led at halftime.
  • Their leads didn’t mysteriously evaporate in the 3rd quarter, however. The Colts led through three quarters in 9 games.
  • The 2017 Colts are the ONLY team in the last 20 years to hold a lead entering the 4th quarter in at least 9 games, but win no more than 4 games.
  • Last year, 25 of 32 teams lost no more than one game when leading entering the 4th quarter, and roughly 35% of them won every game. Six teams lost two games when entering the 4th quarter with a lead. The only team to lose more than twice was the Colts, who lost FIVE times.
  • Last year, a total of five teams held a lead entering the 4th quarter in 9 games. Every team posted a winning record (aside from the Colts):
    • Jaguars (10 wins)
    • Chiefs (10 wins)
    • Falcons (10 wins)
    • Chargers (9 wins)
    • Colts (4 wins)
  • Last year, the only teams to post 4 or fewer wins on the year led after the 4th quarter in an avg of only 2.7 games (aside from the Colts):
    • Browns (0 wins, led after 3Q in 1 gm)
    • Giants (3 wins, led after 3Q in 4 gms)
    • Texans (4 wins, led after 3Q in 3 gms)
    • Colts (4 wins, led after 3Q in 9 gms)
  • In their first 11 games of the season, the Colts led entering the 4th quarter in 8 games. The only two teams that led entering the 4th quarter more than 8 games (through 11) were the Super Bowl Champion Eagles and the Super Bowl runner-up Patriots.
  • Although they led entering the 4th quarter in 8 of their first 11 games, the Colts did not start the year 8-3. Instead they won only 3 of these games. They lost 5 games by blowing leads in the 4th quarter, to drop to 3-8 on the season.
  • The 2017 Colts led by one-score entering the 4th quarter in 6 games, but won just one of six. They were down by one-score entering the 4th quarter in 2 games and lost both. They were 1-7 in games within one-score entering the 4th quarter, despite leading 6 of these games entering the 4th.

Clearly the 2017 Colts were not a great team. But terrible teams are unable to consistently build leads into halftime and into the 4th quarter, like the Colts were routinely able to do. “Something” happened in the 4th quarter to cause such disastrous results. Let’s dive deeply into the analytics of their play calling and decision making to understand WHAT they did, WHY they did it, and HOW it affected the team. Note that I had planned to save this for my 2018 Football Preview book, which I’m writing and will publish in late-June. But this was so incredibly breathtaking, I thought it needed to stand on its own:

ON FIRST DOWN

  • On 1st down in the 4th quarter, if a team is in a one-score game, they run the ball 53% of the time. The Colts ran the ball 64% of the time, 3rd most in the NFL. This despite the fact that on 1st down runs they recorded just a 35% success rate (2.5 YPC), while they were successful on 53% of their passes with 7.8 YPA.
  • In these 4th quarter runs, an older Frank Gore posted just a 30% success rate. A younger, fresher Marlon Mack recorded a 57% success rate, but received a third of the carries that Gore received. As the data shows, potentially due to overuse and wear & tear, Gore was clearly less fresh than Mack, but was still used 3 times more often.
  • On 1st down in the 4th quarter, when winning by 1 score, the Colts ran the ball 79% of the time and recorded a 42% success rate on these runs (2.6 YPC). However, on their passes, they recorded a 100% success rate with 22.0 YPA.
  • On 1st down in the 4th quarter, when winning by 1 score, the Colts used 11 personnel (3 WRs) 30% of the time. The other 70%, they were in 12 or 13 personnel (1-2 WRs).
  • If they had less than 3 WRs, they went 100% run, posting 2.4 YPC and a 38% success rate.
  • Here begins a theme of predictability. A huge key to winning in the NFL is being unpredictable. If the opponent knows your tendencies, you are waging an uphill battle. Especially if your “tendencies” are actually 100% “tells”.
  • How about 1st downs when losing? On 1st down in the 4th quarter, if a team is losing they pass the ball 73% of the time on avg (27% run). But the Colts were 43% run, 2nd most in the NFL. The only team with a greater run rate was the Rams, but they were successful on 57% of these runs. The Colts were successful on just 32%.

ON SECOND DOWN

  • On 2nd down in the 4th quarter when winning, if the Colts did not use 11 personnel (3 WRs) they went 100% run. These runs averaged just 1.7 YPC.

ON BOTH EARLY DOWNS

  • Combining 1st and 2nd down in the 4th quarter, in a one-score game, the Colts ran the ball on 34 of 40 (85%) plays from non-11 personnel (3 WRs). They averaged just 2.0 YPC and a 29% success rate.
  • When not using 3 WRs, they used 12 personnel (1 RB, 2 TEs, 2 WRs) on 95% of their plays, the highest rate in the league. Their 85% run rate when in non-11 personnel was 3rd highest in the league.
  • In the 4Q when playing with a lead, the Colts were the only team in the NFL to NEVER pass unless they had 3 WRs on the field on early downs. If they had a lead & anything less than 3 WRs on the field, they ran 100% of the time. They avg’d 2.1 YPC. There was ZERO threat to pass.
  • Bottom line: the 2017 Colts were the most predictable early down offense in the NFL in the 4th quarter of one-score games. But it gets worse…

ON THIRD DOWN

  • On 3rd down with a 4th quarter lead, once again the Colts were 100% run unless they were in 11 personnel (3 WRs). These runs were so predictable, the Colts posted a 0% success rate and they gained an average of 0 YPC on these runs.

COMBINING EVERY DOWN

  • Combining every down in the 4th quarter, if the Colts were leading, they went 100% run unless they lined up in 11 personnel with 3 WRs. They were the only team in the NFL to go 100% run when fewer than 3 WRs were on the field. With these predictable runs, they gained just 1.9 YPC and recorded a 38% success rate.

OTHER PROBLEMS

  • 4th quarter predictability led to inefficient rushing, which severely hampered the Colts ability to win those 9 games they led entering the 4th quarter. But rushing alone wasn’t the sole cause of their horrible 4th quarter results.
  • The Colts were the only team with zero passing TDs and 2 interceptions. No other team posted 2 interceptions while nursing a one-score 4th quarter lead.
  • In the 4th quarter when leading, the Colts pass efficiency ranked 28th in the league, with only 29% of pass plays grading as successful (avg was 43%). Outside of the Colts 40-yard line, their passing success rate dropped to 21% (avg was 44%).
  • Why were the Colts so bad when passing with a 4th quarter lead? First, we need to understand the Colts were primarily a 11 personnel team when passing, which means 3 WRs. When passing, they used 3 WRs approximately 79% of the time and 2 or fewer WRs 21% of the time.
  • The Colts were substantially more efficient when passing from 2 or fewer WR sets.
    • When using 3 WRs on the season, they were successful on 43% of passes, delivered an 82 rating and averaged 6.8 YPA.
    • When using 2 or fewer WRs, they delivered a 52% success rate, with a 91 rating and 7.8 YPA.
  • The Colts were 1 of only 6 teams to post a sub-45% success rate with 3+ WRs and over-50% success rate with 2 or less WRs.
  • The Colts were extremely successful when passing with 2 or fewer WRs with a lead in the first 3 quarters, recording an incredibly strong 58% success rate on those passes. That rate was 8% better than the NFL average.
  • But for whatever reason, when leading in the 4th quarter the Colts NEVER attempted a pass using 2 or fewer WRs. They only used 3+ WRs. And on these 3+ WR attempts, they recorded a 33% success rate, 6.9 YPA and a 46.2 rating.

The Colts were completely predictable in a number of ways in the 4th quarter. They worked against themselves. They refused to pass out of their most successful personnel groupings for passing. Their predictability in rushing led to inefficiency, which caused hard-earned leads through 3 quarters to slip away in the fourth.

When their leads slipped away, why didn’t Colts come back any time they were trailing by close margins in the 4th quarter?

  • When down one score in the 4th quarter, the NFL average is 64% pass on early downs. When they do run the ball, NFL average is 4.5 YPC and a 49% success rate.
  • When down one score in the 4Q, the Colts ran the ball 10% more than average on early downs. This would only make sense if they were phenomenal when running. But they averaged just 1.9 YPC and posted a 19% success rate. Both were the WORST of any team in the NFL.
  • Meanwhile, on early down passes when down one score in the 4Q, the Colts averaged a 53% success rate (well above NFL average of 47%) and they posted a 104 passer rating (well above the NFL average of 80). Choosing to run 10% more than average & sacrificing such value was unwise.

Here is a look at the overall play calling from the Colts in 2017, courtesy of Sharp Football Stats:

Colts Playcalling

 

Over the course of the entire game, it’s evident the Colts were far too predictable, and featured substantial amounts (see middle column) of

  • Frank Gore on 1st down
  • Frank Gore on 2nd down
  • Hope TY Hilton bails them out on 3rd down

Using this visual data, the amount of “red” in the middle column (representing unsuccessful plays) jumps out at you. Meanwhile, the amount of green for those same down & distance situations in the far right column represents the Colts most successful plays. Clearly they had lots of other possibilities apart from riding Frank Gore so frequently on early downs that weren’t explored enough.

The sad part is, the Colts probably didn’t even realize or measure the impact of their 4th quarter play calling. It was far worse than they could have imagined. My guess is they had no idea that they NEVER passed while maintaining a 4th qtr lead without 3 WRs on the field.

They likewise probably had no idea that the only teams to enter the 4Q with a lead more often than themselves (through week 12) faced each other in this year’s Super Bowl: the Eagles & Patriots.

It’s unfortunate to sit back now and realize that many of these 4th quarter leads which became losses were avoidable with stronger attention to detail and a better focus on analytics.

Often time fans and media are quick to place blame on players making a mistake on the field, without realizing the play call wasn’t optimal to begin with, and majority blame should shift elsewhere. Understanding responsibility for the error is essential to correcting it and ensuring it doesn’t become repetitive. Because repetition leads to habit-forming behavior. And unfortunately, that is exactly what happened to the Colts play calling.

When placed into certain situations (leading in the 4th quarter) the Colts changed their strategy and style of play which earned them the lead through the first three quarters. They played tighter. They played predictably. They played not to lose. They refused to use their optimal play calls. They forced their quarterback (experienced as he was) into predictable passing situations and allowed the defense to attack, knowing what to expect.

Reviewing this in hindsight is certainly infuriating. I will be doing similar dives into the 31 other teams for my 2018 Football Preview book, out in late-June. I’m guessing none will be quite as eye-opening as what the 2017 Colts did to themselves in the fourth quarter and the monumental impact it had on their final record, but we shall see.

For the 2017 Colts, this is what happens when a team does not pay attention to detail. Details they would only know if they incorporated more analytics. Analytics isn’t a dirty buzz word. Teams have been winning Super Bowls for decades using analytics, such as Bill Walsh’s 49ers. Analytics is simply the use of math or statistics to gain valuable knowledge from data. If you add up the rushing yards of the Colts from 1-2 WR sets when leading in the 4th quarter, and divide by rushing attempts, and realize that these runs are totally inefficient, you’re essentially using analytics. Sounds far less scary and more basic, and teams to accept this level of detail with open arms and incorporate it into their arsenal of weaponry as they try to improve and put the best product on the field which gives them the best chance to succeed.

Once more teams start self-scouting in this manner, they will avoid more pitfalls like the 2017 Colts faced. The impact it had on their season and final record was profound. Additionally, this level of analysis is also able to find holes in opponents on a weekly basis. Such as why I believed the Patriots would have success throwing on the #1 pass defense of the Jaguars in the AFC Championship, or why I believed the Eagles would have immense success running on the Patriots in the Super Bowl.

With the Eagles winning the Super Bowl this year, seeing a ton of success thanks to incorporating far more analytics into their team than most, I believe we’re turning the corner. Teams that do not use more analytics to self-scout and identify their own strengths and weaknesses they weren’t otherwise aware of, or to scout their opposition, will soon fall behind. Way behind. “Analytics” are not a robot overlord, send out of the matrix to tell a coach what to do, expecting subservience. Intelligently incorporating analytics will not paralyze a team nore make them too smart for their own good.

Working with experts who can share analytical insights tactfully, while allowing the coaching staff to take what they want from the information as part of their own final decision making process, will no longer be the path less traveled in the NFL. It is now, will be in the future, and truthfully, has been for some decades, the easiest way to win in the NFL.