John Harbaugh: We understand the math on fourth downs, going for two

Getty Images

Ravens coach John Harbaugh is standing by his decisions in Sunday’s loss to the Chiefs, decisions that included going for it on fourth down four times, and going for two after three of the Ravens’ four touchdowns.

Harbaugh said he talks to the Ravens’ analytics staff and formulates a plan for each individual opponent, and against the Chiefs, the Ravens wanted to be aggressive about trying to make the most out of each possession, which means trying to score eight points rather than end possessions with a kick.

“I have a good understanding of the numbers,” Harbaugh said. “It’s very detailed and well thought out. I think I was pretty clear about it last night, but we’re standing by our decisions. Our decisions gave us the best chance to win the game in that particular game. These are not league average choices. These are determined by this game and for this game specifically, in that venue, weather is even factored into it. There’s a lot of factors that go into it that are mathematically calculated. And that’s why we did it. It wasn’t a field position game. It was a possession game. And making the most of each possession was what counted and that’s what we were attempting to do and for the most part we did a really good job of it.”

When the Ravens scored a touchdown early in the fourth quarter to cut the Chiefs’ lead to 11 points, they went for two and failed. Many observers questioned that decision, but Harbaugh defended it, saying the math is clear that cutting the lead to nine points would have significantly improved their chance of winning.

“Getting it to nine gives you a much better chance of winning,” Harbaugh said. “You still have a chance to do that with the second two. And if for some reason they happen to kick a field goal or score a touchdown, it also enhances your odds. So while you may think getting to 10 is the thing to do, it’s the thing to do if you want to go to overtime. It’s not the thing to do if you want to win the game in regulation, and that’s what we were trying to do.”

Harbaugh said that while he listens to the analytics staff, he listens more to his own intuitions based on a life spent in football. In this cases, Harbaugh says, the analytics people were in harmony with his gut.

“The analytics guys will tell you I don’t follow the analytics nearly enough,” Harbaugh said. “They’ll tell you I go by my gut way more than I go by the analytics, and I do. Because the flow of the game, the feel of the game, situations you’ve been in . . . all those things are something, as a coach, you have a real sense for. So I’ll go against the analytics a lot more than I’ll go with it, in terms of 50/50 close calls. But in a game like that, those were definitely decisions that gave us the best chance to win, and put us in the best position to win the game, no question. If we hadn’t made those decisions, especially the fourth-down decisions, we wouldn’t have been within a score at the end of the game. Period.”

Analytics experts would agree with that: The math is clear that NFL coaches should go for it on fourth down more often than they do, and that going for two makes sense at times other than the obvious cases of a team trailing by eight points scoring a touchdown to cut the deficit to two points late in the game. The Ravens lost, but the analytics experts would agree with Harbaugh that he made the decisions that gave his team the best chance to win.

68 responses to “John Harbaugh: We understand the math on fourth downs, going for two

  1. If the probability of making two points is 50 percent and the probability of making one point is 90 percent, you should go for two every time.

  2. “If the probability of making two points is 50 percent and the probability of making one point is 90 percent, you should go for two every time.”

    Umm… what if you are down by 7 with 5 seconds left? Gonna go for two every time? Maybe sometimes, but not every time. This isn’t accounting or a mutual fund where we’re looking at returns over two decades. It’s a football game and much more finite.

  3. His decision to go for 2 in the 4th to possibly make it a 9-point game or leave it an 11-point game PROVES his lack of understanding of the math, despite whatever gibberish he offers to try to ‘splain away his decision.

  4. Just simplify it and ask yourself what Bill Belichick would do.

    He would have went with the field goal.

    Granted, the kicker would’ve missed it.

    But Bill would go with the field goal.

  5. viffty says:
    September 23, 2019 at 4:05 pm
    Once again Harbaugh tried to be the smartest guy in the room and failed.

    73 19 Rate This

    ————————

    He thinks he’s on BB;s level. It’s hilarious.

  6. rewan nayshen says:
    September 23, 2019 at 4:17 pm
    If the probability of making two points is 50 percent and the probability of making one point is 90 percent, you should go for two every time.

    ————-

    That’s a common misconception.

    Probability doesn’t work that way. If you take something that is truly 50% like a coin flip, you could easily flip heads 4 times in a row. That could represent 4 straight fails on a 2 point convert in a game. The probability doesn’t play out over such a small set of events. A team could miss most of their 2 point tries on a season and it still wouldn’t be a big enough set of events to ensure a 50% outcome. That doesn’t even consider that every single play has it’s own variables. What is the score in the game, what is the weather like, what players are injured, what plays have been working, is the defense tired, etc.

    A kick is much more predictable on an individual play because the probability is so high.

  7. Harbaugh is one of the best coaches in the NFL, and there are a lot of good ones. He’s been a big winner, and that’s because he coaches to win. You can’t coach scared when you play the Chiefs. John Harbaugh knew more about coaching by the time he was 15 years old, than most of us will ever know.

  8. “If the probability of making two points is 50 percent and the probability of making one point is 90 percent, you should go for two every time.”

    Without even taking you to task for the gross assumptions and incredibly small sample size of situations that go into this math – if your objective is to ignore game situation and simply maximize the number of points you score over an entire season, then by all means do this.

  9. “what if you are down by 7 with 5 seconds left? Gonna go for two every time?”

    You could still miss the extra point or lose in OT. Some would say playing to win or playing not to lose.

  10. Harbaugh is a good football coach…and deserves the benefit of the doubt…but please just do what Belichick does and…#SayLess

  11. One issue with his assessment of the odds is that they are averages. His team has the best FG kicker, and one of the best punters, in the NFL. They can get results more consistently than the average (made FG tries, punting placement). That has to be factored in.

    Harbaugh is an excellent coach but he can get too clever for his own good at times.

  12. What JH is failing to acknowledge here is that his adherence to the analytics cost him 2 points already in the game. Had he been kicking XPs, he’d likely have only been down 9 at that point, with a chance to go down by 8 (one score) with a kicked XP. Analytics may be great and all, but he left 3 points on the field because of it.

  13. I understood what Jon Harbaugh was doing trying to convert on 4th downs .. Lamar Jackson is explosive and scary to defend at the NFL level too …

    but I don’t like to see coaches leave points on the table, especially at the beginning of a game … take em when you can and use the 2 point option if you need to … at the end of the game.

  14. Coaches who use this logic and go for two all the time seem to fail 2-3 times in a row more often than not.

  15. The Ravens kicked the extra point after 1st TD but decided to go for 2 because of the offside penalty on KC and failed. Why did they not go for 2 before the penalty? It seems to me that they were chasing those missed extra points all game long. All this analytics and stuff does not make lot of sense in this situation, the Steelers tried going for 2 every time for a season or two and abandoned it. There aren’t many plays from the 2 yd line, you save those plays when you really need it like in a must win game.

  16. FratrickMahoes says:
    September 23, 2019 at 4:30 pm
    I don’t even make this risky of play calls in Madden.

    70 10 Rate This

    Great post buddy!!!!!

  17. tylawspick6 says:
    September 23, 2019 at 4:59 pm
    viffty says:
    September 23, 2019 at 4:05 pm
    Once again Harbaugh tried to be the smartest guy in the room and failed.

    73 19 Rate This

    ————————

    He thinks he’s on BB;s level. It’s hilarious.

    ————————
    Agreed he’s not even close, BB has punked him to many times so cant deny it. But Im going to side with Harbaugh on this one. Against that opponent that was very much a good strategy. Had the conversions worked better that would have changed the game, likely change the outcome. Despite his misadventures with BB (and its nit like he is alone there) he us still on the overall one of the leagues better coaches.

  18. Didn’t like all the 2pt attempts – thought it got a bit ridiculous (0 for 3) – but the bigger problem was the offensive game plan – they should have run it down KC’s throat and controlled the clock – Lamar is improving but he’s not Mahomes and they asked him to do too much, especially in the first half

  19. Ingram and Edwards were having a field day on the ground, i wonder why you would throw the ball so many times. But we will see if Harbs hypothesis works. Go Ravens!

  20. Yes, cutting the lead from 11 to 9, versus 10, gives you a better chance of winning, but potentially eschews a legitimate shot at OT. I personally appreciate the aggressive approach. If it works, you likely win. But it didn’t work. And regardless of which choice you make, lack of success brings second guessing. The Ravens consistently went for it. It was their game plan. The Chiefs just stopped them. Credit the Chief defense for outstanding work. Now he set the tone early after taking the 1 point off the board following the first TD to try for 2 and failing. In that case he wasn’t trying for 2 versus 1. He was surrendering an established 1 for a chance at a 2. And even at a yard closer, he may be alone in this thoughts on the value of that exchange.

  21. Harbaugh is a dang good coach that probably 25+ teams would love to have. Second guess him all you want. At least he’s owning up to it and standing by his decision. They lost, on the road in bad weather, to a very good team by just a few points. The Ravens are contenders this year, much in part to Harbaugh and his coaching.

  22. he has one of the best kickers around, so if he doesn t plan on kicking then just trade the guy and hire a loser. get a draft pick and save money.

  23. His logical on going for 4th down is something that has to be calculated on the yardage and other factors but okay – just e aware that ypir kicker is going to boot 99.99% within 40 yards in – I could see not wanting to leave a short field for KC if they miss but that’s what punting is for also. you can pin them down on the 1.

    As for the 2-point conversion, it makes no sense early especially against a team that can score in 7 seconds. You trade 1 sure poi t for ZERO and then you have to do it after a second TD to try and catch up but now instead of being almost certainly tied, they are down by a FG AFTER scoring 2 TD’s – not smart. It’s a strategy that can work on the Jets as if you get 2 TD’s and 2 points, there’s no way they can get3 scores by KC? They can score 28 points in 9 minutes – why are you screwing with math?

  24. tylawspick6 says:
    September 23, 2019 at 4:59 pm
    viffty says:
    September 23, 2019 at 4:05 pm
    Once again Harbaugh tried to be the smartest guy in the room and failed.

    73 19 Rate This

    ————————

    He thinks he’s on BB;s level. It’s hilarious.

    ————————————————-
    The same Harbaugh that beat the Pats three ties in the postseason?

  25. Been watching(and playing)football seriously for about 40 years now. The playing part is long in the past but I still love the game.

    I have never seen that before as far as I can remember. I know the 2-pt conversion is relatively new in the NFL but it’s been around a long time in college so we’ve seen a lot of situations over the years.

    I think the reason I had never seen it before is because no other coach in that 40 years thought it made any sense. The math doesn’t add up.

    I’m a critical thinker and over-analyze everything(drives the wife crazy)and I sat there for ten minutes after that call trying to understand the logic behind it and all thinking about it did was frustrate me.

    And I’m a Seahawks fan so it wasn’t a Raven’s fan thing…it was more of a football fan wondering wth thing. It was all wrong…it was like a Fantasy Football player who didn’t really understand the game was in charge for a second. It made me cringe.

    It was like when you see a bumper-sticker on the back of a nice car but the person put it on all crooked and off-center. The whole thing just felt wrong and it really bothered me for awhile afterwards. My wife smiled and nodded as I spouted off about it, thank god for her. She probably tuned me out but it really helped me get over it. I was actually a little upset about it for some reason.

    BTW, His explanation here didn’t help me understand.

  26. Bottom line, you are either down by a TD and FG or you are down by TWO TD’s….

    The latter is much harder to achieve at the end of a game. Getting a FG at the end to tie happens all the time.

  27. Say what you will but i would sleep like a baby at night really liking the Chiefs chances in winning a Superbowl if John Harbaugh was their Head-Coach, You saw Patrick Mahomes upset and puzzle why Andy Reid is calling timeouts just before the half with tons of time left on the clock to kick Field goals, rather than go for Touchdowns to really put teams away so you have zero issues in the dying seconds trying to get a first down in order to close out the game,Especially when your Defense is the WEAK LINK.

  28. “It’s not something that anybody’s ever done before.” John Harbaugh
    “It’s a play that we thought would work.” Former HC of the NYJ
    “Maybe those guys gotta study the rule book and figure it out.” — the QB of the NEP
    “Chess is life.” Bobby Fischer

  29. Well I’m not surprised to see so many Patriots fans tell us their coach is smarter. I’m also not surprised to see so many people who think they are smarter than analytics despite the closing paragraph (if you bothered to read it) that says analytics experts agree.

    I have no problem with these decisions. The Ravens have an offense that is dynamic and unpredictable. It scored 28 points which is usually enough to win a game. The defense however gave up 33 points and 374 passing yards. That’s the real problem. Jackson and the offense are looking good. The defense, especially the secondary despite having so much talent, didn’t hold up and the defense couldn’t generate much of a pass rush which put more pressure on the secondary.

  30. Harbaugh is only NFL coach to go into Foxboro and beat BB and the Patriots in the playoffs twice. It should have been three times if Lee Evan hangs on to a perfectly thrown TD pass from Flacco, and Cundiff does not miss a chip shot field goal. BB benefits from having the GOAT QB, tremendous personnel and playing in a cupcake division for the last 20 years. I get what Harbs was trying to do. He knew there was no way his defense was going to keep Mahomes under 30, so he took his chances and tried everything possible to break 30 to increase Baltimore’s win percentage. He played to win and it almost worked – KC best team in football. No shame in that loss or strategy. It is actually the same data analytics strategy the Eagles employed in the Super Bowl to beat the Patriots. Remember when they went for it on 4th & goal at end of first half when they were already up by 3 (the “Philly-Philly” play where Foles caught the TD pass) — everyone wondered why they did not take the sure chipshot 3 to go up 6 going into halftime. By playing the percentages and going for the TD (and making it) — their win percentage shot up nearly an additional 20% by going up by 10 over going up 6 – (a +4 swing in points). Given their opponent Pederson knew he had to score as many points as possible (it was not going to be a low scoring game like last year’s Pats/Rams) and thus worth the calculated risk against THAT opponent in that game. Pederson went for it a lot in that game as I recall. If he had a defense like Wade Phillips had with the Rams- he may have kicked the FG. A lot of scenarios are factored in the data analytics. Hate to burst bubbles – but the Patriots also employ a data analytics firm like The Ravens and Eagles.

  31. Harbaugh is exactly right. Its amazing that more teams dont do this. If you are the inferior team, you should always play this way. Its actually simple analytics.

    Ravens would probably have about a 40% chance to win in OT. Assume 50% for 2 pt conversion rate.

    You score a TD to get down 8 (before the conversion). The stupid announcers say its crazy to go for 2. They are wrong. Lets look at the possible outcomes (assuming you score another TD- which is required for the issue not to be moot)

    You have a 50% chance of getting to the point where a touchdown basically wins the game if you go for 2.

    You have a 25% chance of not getting to overtime (miss both=25%) chance.

    You have a 25% chance you miss the first and make the second- this going to OT.

    So, you win 50% in regulation and 10% in OT (25% chance of going to OT times 40% chance of winning if you go to OT) for a total win chance of 60% if you get the TD required

    You kick two extra points, even giving it 100% chance of making both- you only win 40% of the time in OT.

    Why is it better to not go for two?

    I salute Harbaugh for using some sense and not being afraid to actually use math instead of being a “football guy” and actually maximizing his chance to win.

    Its amazing how many people here can not figure out even basic probability.

  32. In short, you are ordered to do what Vegas tells you to do, that is all. Before anyone bashes this comment, just know that these point spreads play a huge deal and they are artistic how to manage games and put up certain numbers. A couple points can turn out to be huge swings in terms of the spread. I love football just like anyone here but I know this is a controlled sport.

  33. ballreaper says:
    September 23, 2019 at 10:37 pm
    Say what you will but i would sleep like a baby at night really liking the Chiefs chances in winning a Superbowl if John Harbaugh was their Head-Coach, You saw Patrick Mahomes upset and puzzle why Andy Reid is calling timeouts just before the half with tons of time left on the clock to kick Field goals, rather than go for Touchdowns to really put teams away so you have zero issues in the dying seconds trying to get a first down in order to close out the game,Especially when your Defense is the WEAK LINK.

    —————

    No question that Reid needs to be more aggressive. In the AFCCG he had 11 seconds left and he took a FG to tie instead of taking one more shot to the endzone with Mahomes to ice the game. He was worried about a sack or a pick in the endzone I guess, but when you have Mahomes you have to give him a shot. Then he lost the OT coin flip and Reid was all upset in the offseason that Mahomes didn’t get on the field. Reid has always been bad situationally at the end of games and he still is.

  34. ballreaper says:
    September 23, 2019 at 10:37 pm
    Say what you will but i would sleep like a baby at night really liking the Chiefs chances in winning a Superbowl if John Harbaugh was their Head-Coach
    ————————————–
    If my only choices are Reid or Harbaugh, I take Reid every time. Yes Harbaugh lucked into a stellar D that won him a SB but he is not on Reids level when you speak of coaching ability. If he was, he would’ve started the game running the ball and he would’ve run until the Chiefs could stop it. Which they clearly couldn’t.

  35. lifelongwhodat says:
    September 24, 2019 at 12:23 am
    Harbaugh is exactly right. Its amazing that more teams dont do this. If you are the inferior team, you should always play this way. Its actually simple analytics.

    Ravens would probably have about a 40% chance to win in OT. Assume 50% for 2 pt conversion rate.

    You score a TD to get down 8 (before the conversion). The stupid announcers say its crazy to go for 2. They are wrong. Lets look at the possible outcomes (assuming you score another TD- which is required for the issue not to be moot)

    You have a 50% chance of getting to the point where a touchdown basically wins the game if you go for 2.

    You have a 25% chance of not getting to overtime (miss both=25%) chance.

    You have a 25% chance you miss the first and make the second- this going to OT.

    So, you win 50% in regulation and 10% in OT (25% chance of going to OT times 40% chance of winning if you go to OT) for a total win chance of 60% if you get the TD required

    You kick two extra points, even giving it 100% chance of making both- you only win 40% of the time in OT.

    Why is it better to not go for two?

    I salute Harbaugh for using some sense and not being afraid to actually use math instead of being a “football guy” and actually maximizing his chance to win.

    Its amazing how many people here can not figure out even basic probability.

    ——————

    From a pure mathematical probability, yes that is true.

    However, the law of averages cannot play out over two occurrences, so you are really leaving the game up to random chance going for two based solely on the math. If Harbaugh was considering many other situational variables in the calculation, then it may have been sensible.

    Also, mathematical probability assumes events are independent, and in a football game they are not. Being within 7 may have an impact on he opposing teams offensive play calling for example. Every play has an impact on future plays that needs to be factored in.

  36. restonraven says:
    September 24, 2019 at 12:19 am
    Harbaugh is only NFL coach to go into Foxboro and beat BB and the Patriots in the playoffs twice. It should have been three times if Lee Evan hangs on to a perfectly thrown TD pass from Flacco, and Cundiff does not miss a chip shot field goal. BB benefits from having the GOAT QB, tremendous personnel and playing in a cupcake division for the last 20 years. I get what Harbs was trying to do. He knew there was no way his defense was going to keep Mahomes under 30, so he took his chances and tried everything possible to break 30 to increase Baltimore’s win percentage. He played to win and it almost worked – KC best team in football. No shame in that loss or strategy. It is actually the same data analytics strategy the Eagles employed in the Super Bowl to beat the Patriots. Remember when they went for it on 4th & goal at end of first half when they were already up by 3 (the “Philly-Philly” play where Foles caught the TD pass) — everyone wondered why they did not take the sure chipshot 3 to go up 6 going into halftime. By playing the percentages and going for the TD (and making it) — their win percentage shot up nearly an additional 20% by going up by 10 over going up 6 – (a +4 swing in points). Given their opponent Pederson knew he had to score as many points as possible (it was not going to be a low scoring game like last year’s Pats/Rams) and thus worth the calculated risk against THAT opponent in that game. Pederson went for it a lot in that game as I recall. If he had a defense like Wade Phillips had with the Rams- he may have kicked the FG. A lot of scenarios are factored in the data analytics. Hate to burst bubbles – but the Patriots also employ a data analytics firm like The Ravens and Eagles.

    1 4 Rate This

    ———————-

    Umm, ok, in 2009 NE wasn’t really a good playoff team and Welker blew his leg out the week before.

    NE was transitioning away from the Old Guard and trying to get young into the Lockout and to load up on draft pucks. They sold off Seymour, Vrabel, etc, and HOFers had just retired or were really old.

    I mean, my goodness, talk about overselling it. 2009 was arguably NE’s worst playoff team, and in fact, it was.

    If this is your level of being impressed, no wonder why you are so impressed with Harbaugh now, 10 years later.

    Yikes!

    NE will slaughter your team down there later in the year. You’re never as good as think you are, which is why you’re constantly a .500 team the last 4 years.

  37. Two points: (1) the more you go for two the better you get at it which gives you an advantage and (2) if you have the confidence that you can make over 50% than you will come out ahead in an individual game. Going for two is sound. Kickers don’t make 100% either.

  38. “If the probability of making two points is 50 percent and the probability of making one point is 90 percent, you should go for two every time.”

    Umm… what if you are down by 7 with 5 seconds left? Gonna go for two every time? Maybe sometimes, but not every time. This isn’t accounting or a mutual fund where we’re looking at returns over two decades. It’s a football game and much more finite.
    ******************************8

    According to Harbaugh, yes, he would go for 2 to win, because he said he’s trying to win, not go to OT. Let’s see if he actually does this when presented that opportunity, i’d guess that he kicks and goes to OT.

  39. See…this is when mathematicians and fantasy football players take over from the people that have watched/played football their entire lives and understand the game better.

    Simply put, the football is not round. You cannot predict which way it bounces.

  40. The Chief’s were already living rent free in Harbaugh’s head before the game even began. He figured the only way he was going to have a chance of keeping up with Mahome’s was to grab every point he could, when he could. It didn’t really cost them the game though, because it really wasn’t as close as the final score suggests. A little prevent D from the Chief’s in the 4th made things a bit closer, but that was it.

  41. This guy is as avg as his brother except he has a super bowl. Not impressed with either one and Jim bo is up to his eyeballs is average players he recruited at Mich.

  42. If you are playing a high octane offense like KC, it makes sense to score TDs instead of FGs and 8 points rather than 7. Not to mention KC has one of the worst defenses in the NFL. Just because Baltimore lost doesn’t mean it was a bad strategy. They covered the spread which by definition means they did better than expectations.

Leave a Reply

You must be logged in to leave a comment. Not a member? Register now!

This site uses Akismet to reduce spam. Learn how your comment data is processed.