Tomlin made the stupidest coaching decision ever (1 Viewer)

[rick6668]

I knew footballcommentary.com would "comment" on this. Looks like, statistically, it's about the same whether you go for it or kick the XP.

A more interesting situation arose after the Steelers scored a touchdown with 10:25 left in regulation to pull within five points prior to the try. The two-point conversion apparrently succeeded, but Pittsburgh was flagged for holding, resulting in a re-try from the 12-yard line. The Steelers once again went for two, but failed.

According to the Chart, Pittsburgh should go for two in that situation if the probability of success exceeds about 0.19. That's close to the actual probability of scoring from the 12-yard line. So ostensibly, it's a close call. However, the Chart doesn't take into account that if the Steelers kick the extra point, it's a 30-yard attempt rather than the normal 20-yard attempt, and the success probability is only about 90% rather than 99%. This reduces the required probability of success on a two-point conversion from 0.19 to 0.17, and makes going for two slightly preferred despite the penalty. However, there is very little difference in win probability regardless of Pittsburgh's choice.

Great post. I would love to see how they figure that you only need to convert 17% of the time to justify going for the two point conversion. I don't believe it. But still it is interesting to see somebody seriously defending Tomlin's decision. I know it would have been a serious mistake to kick the extra point from the two yard line. If they could convince me that Tomlin was right I would take great pleasure to eat some crow. If it is even close, like 30%, I suppose it wouldn't be the stupidest decision ever.

OK I have read the article at footballcommentary.com and I am ready to eat crow. I am shocked, but their methodology looks very solid. I doubt this is the chart that Tomlin uses, but maybe it is. I am shocked, simply shocked, that going for the 2 was right, but now I believe. I take back every insult I hurled at Tomlin. Congrats to KnowledgeReignsSupreme for showing me the way.

I read this too, I'd still like to know how they got the probabilty of scoring a TD from the 12 yard line as .19.That's basically one out of every 5 times. Maybe that's correct, just like to see what led them to that number.

 

[Synthesizer]

I knew footballcommentary.com would "comment" on this. Looks like, statistically, it's about the same whether you go for it or kick the XP.

A more interesting situation arose after the Steelers scored a touchdown with 10:25 left in regulation to pull within five points prior to the try. The two-point conversion apparrently succeeded, but Pittsburgh was flagged for holding, resulting in a re-try from the 12-yard line. The Steelers once again went for two, but failed.

According to the Chart, Pittsburgh should go for two in that situation if the probability of success exceeds about 0.19. That's close to the actual probability of scoring from the 12-yard line. So ostensibly, it's a close call. However, the Chart doesn't take into account that if the Steelers kick the extra point, it's a 30-yard attempt rather than the normal 20-yard attempt, and the success probability is only about 90% rather than 99%. This reduces the required probability of success on a two-point conversion from 0.19 to 0.17, and makes going for two slightly preferred despite the penalty. However, there is very little difference in win probability regardless of Pittsburgh's choice.

Great post. I would love to see how they figure that you only need to convert 17% of the time to justify going for the two point conversion. I don't believe it. But still it is interesting to see somebody seriously defending Tomlin's decision. I know it would have been a serious mistake to kick the extra point from the two yard line. If they could convince me that Tomlin was right I would take great pleasure to eat some crow. If it is even close, like 30%, I suppose it wouldn't be the stupidest decision ever.

OK I have read the article at footballcommentary.com and I am ready to eat crow. I am shocked, but their methodology looks very solid. I doubt this is the chart that Tomlin uses, but maybe it is. I am shocked, simply shocked, that going for the 2 was right, but now I believe. I take back every insult I hurled at Tomlin. Congrats to KnowledgeReignsSupreme for showing me the way.

I read this too, I'd still like to know how they got the probabilty of scoring a TD from the 12 yard line as .19.That's basically one out of every 5 times. Maybe that's correct, just like to see what led them to that number.

I know exactly how they got to it..... fuzzy math.

 

[J R]

I knew footballcommentary.com would "comment" on this. Looks like, statistically, it's about the same whether you go for it or kick the XP.

A more interesting situation arose after the Steelers scored a touchdown with 10:25 left in regulation to pull within five points prior to the try. The two-point conversion apparrently succeeded, but Pittsburgh was flagged for holding, resulting in a re-try from the 12-yard line. The Steelers once again went for two, but failed.

According to the Chart, Pittsburgh should go for two in that situation if the probability of success exceeds about 0.19. That's close to the actual probability of scoring from the 12-yard line. So ostensibly, it's a close call. However, the Chart doesn't take into account that if the Steelers kick the extra point, it's a 30-yard attempt rather than the normal 20-yard attempt, and the success probability is only about 90% rather than 99%. This reduces the required probability of success on a two-point conversion from 0.19 to 0.17, and makes going for two slightly preferred despite the penalty. However, there is very little difference in win probability regardless of Pittsburgh's choice.

Call me stubborn, but I don't believe this can be correct. First, I find it hard to believe that the chances of success from the 12 yard line are anywhere near 19%. Where are they getting this data from? Surely, it isn't from a rich database of 2-point conversion from the twelve. So it must be some amalgam of other scenarios such as 3rd and 12 conversions. I think this is wrong because converting 3rd and twelve from midfield is certainly much more difficult than converting Tomlin's 2-point conversion because there is far less room to work with. That is, on 3rd and 12 from midfield you can complete a 30 yard pass. A 30 yard pass play was not an option for Pitt in their situation. Secondly, is a 30 yard FG/XP really only 90% likely or is this the likelihood of converting from 30-39 yards?

Your stubborn response sounds correct to me, better than KRS just assuming they're right. But I don't think they're comparing the probability of scoring on that play to the probability of converting a 3rd and 12. They say "That's close to the actual probability of scoring from the 12-yard line." I read that to mean the probability of scoring from the 12 in all scenarios.You'd have to be pretty gullible to fall for their math there. They basically say that the Steelers had the same odds to score whether they were on the 2 or the 12.

 

[pecorino]

the "math" here is completely fictitious. they have no data to go on as i claimed in a previous post.

a local sports guy estimated the chances of converting from the 12-yard line at 3% and, while it was clearly an off-the-cuff guess, it rings way more true than 17%.

 

[KnowledgeReignsSupreme]

the "math" here is completely fictitious. they have no data to go on as i claimed in a previous post.a local sports guy estimated the chances of converting from the 12-yard line at 3% and, while it was clearly an off-the-cuff guess, it rings way more true than 17%.

Oh, I see. So we're going on the word of a local sports guy guessing off the cuff?I understand where you're coming from. It doesn't seem right. It's almost counter intuitive. But that doesn't mean it's wrong.

 

[KnowledgeReignsSupreme]

I knew footballcommentary.com would "comment" on this. Looks like, statistically, it's about the same whether you go for it or kick the XP.

A more interesting situation arose after the Steelers scored a touchdown with 10:25 left in regulation to pull within five points prior to the try. The two-point conversion apparrently succeeded, but Pittsburgh was flagged for holding, resulting in a re-try from the 12-yard line. The Steelers once again went for two, but failed.

According to the Chart, Pittsburgh should go for two in that situation if the probability of success exceeds about 0.19. That's close to the actual probability of scoring from the 12-yard line. So ostensibly, it's a close call. However, the Chart doesn't take into account that if the Steelers kick the extra point, it's a 30-yard attempt rather than the normal 20-yard attempt, and the success probability is only about 90% rather than 99%. This reduces the required probability of success on a two-point conversion from 0.19 to 0.17, and makes going for two slightly preferred despite the penalty. However, there is very little difference in win probability regardless of Pittsburgh's choice.

Call me stubborn, but I don't believe this can be correct. First, I find it hard to believe that the chances of success from the 12 yard line are anywhere near 19%. Where are they getting this data from? Surely, it isn't from a rich database of 2-point conversion from the twelve. So it must be some amalgam of other scenarios such as 3rd and 12 conversions. I think this is wrong because converting 3rd and twelve from midfield is certainly much more difficult than converting Tomlin's 2-point conversion because there is far less room to work with. That is, on 3rd and 12 from midfield you can complete a 30 yard pass. A 30 yard pass play was not an option for Pitt in their situation. Secondly, is a 30 yard FG/XP really only 90% likely or is this the likelihood of converting from 30-39 yards?

Your stubborn response sounds correct to me, better than KRS just assuming they're right. But I don't think they're comparing the probability of scoring on that play to the probability of converting a 3rd and 12. They say "That's close to the actual probability of scoring from the 12-yard line." I read that to mean the probability of scoring from the 12 in all scenarios.You'd have to be pretty gullible to fall for their math there. They basically say that the Steelers had the same odds to score whether they were on the 2 or the 12.

The odds of converting from the 2 are at least 40%. No one has made the claim that the odds of scoring from the 12 are 40%. Please stop doing this.

 

[KnowledgeReignsSupreme]

email sent to footballcommentary for clarification.

 

[J R]

I knew footballcommentary.com would "comment" on this. Looks like, statistically, it's about the same whether you go for it or kick the XP.

A more interesting situation arose after the Steelers scored a touchdown with 10:25 left in regulation to pull within five points prior to the try. The two-point conversion apparrently succeeded, but Pittsburgh was flagged for holding, resulting in a re-try from the 12-yard line. The Steelers once again went for two, but failed.

According to the Chart, Pittsburgh should go for two in that situation if the probability of success exceeds about 0.19. That's close to the actual probability of scoring from the 12-yard line. So ostensibly, it's a close call. However, the Chart doesn't take into account that if the Steelers kick the extra point, it's a 30-yard attempt rather than the normal 20-yard attempt, and the success probability is only about 90% rather than 99%. This reduces the required probability of success on a two-point conversion from 0.19 to 0.17, and makes going for two slightly preferred despite the penalty. However, there is very little difference in win probability regardless of Pittsburgh's choice.

Call me stubborn, but I don't believe this can be correct. First, I find it hard to believe that the chances of success from the 12 yard line are anywhere near 19%. Where are they getting this data from? Surely, it isn't from a rich database of 2-point conversion from the twelve. So it must be some amalgam of other scenarios such as 3rd and 12 conversions. I think this is wrong because converting 3rd and twelve from midfield is certainly much more difficult than converting Tomlin's 2-point conversion because there is far less room to work with. That is, on 3rd and 12 from midfield you can complete a 30 yard pass. A 30 yard pass play was not an option for Pitt in their situation. Secondly, is a 30 yard FG/XP really only 90% likely or is this the likelihood of converting from 30-39 yards?

Your stubborn response sounds correct to me, better than KRS just assuming they're right. But I don't think they're comparing the probability of scoring on that play to the probability of converting a 3rd and 12. They say "That's close to the actual probability of scoring from the 12-yard line." I read that to mean the probability of scoring from the 12 in all scenarios.You'd have to be pretty gullible to fall for their math there. They basically say that the Steelers had the same odds to score whether they were on the 2 or the 12.

The odds of converting from the 2 are at least 40%. No one has made the claim that the odds of scoring from the 12 are 40%. Please stop doing this.

That makes sense to me, KRS.

 

[J R]

OK, this is hardly scientific, but until some real data comes in I thought it might at least be interesting. I count 25 plays that started between the opponents' 10 and 15 in the wild-card games (after taking out FGs, and including the 2-point play in the Steelers-Jags game). 2 out of 25 were TDs, an 8% TD rate. Out of those 25 plays, 5 were 3rd downs, which are similar to 2-point conversions because they are 1-shot propositions, but different in that teams can always settle for first down conversions instead of TDs or kicking FGs or even going for it on 4th down. Teams turned 0 of those 5 into TDs. Teams other than the Titans, who seem unlikely to score in such situations since Young isn't good, were 2 for 20, a 10% rate.

It's a small sample, sure. I'm just having a hard time believing the probability of scoring from the 12 is greater than 19%.

 

[KnowledgeReignsSupreme]

I read this too, I'd still like to know how they got the probabilty of scoring a TD from the 12 yard line as .19.

That's basically one out of every 5 times. Maybe that's correct, just like to see what led them to that number.

email sent to footballcommentary for clarification.

Here is his response:

Thank you for your interest in footballcommentary.com.

I do not have a database with all the 4th-and-goal-from-the-12 plays. (That situation doesn't come up very often, and when it does, teams rarely go for it. So I don't know how much useful data there would be.) Aaron Schatz at Football Outsiders once told me that the likelihood of making a first down (or a touchdown, if it's goal-to-go) is about 31% on 3rd and 10, and 22% on 3rd and 15. This ostensibly suggests somewhere around 27% for the likelihood of making a first down on 3rd and 12, and 4th down should be similar. However, those cases are mostly not goal-to-go situations, which are harder to convert. Also, the quoted percentages include picking up the first down via a defensive penalty, which is less helpful in the situation Pittsburgh faced. So it's definitely appropriate to use a much lower probability of success, which is what I did.

If people have evidence that 20% still significantly overstates Pittsburgh's chances, that's fine, but I think they're missing the point. The conclusion that it doesn't matter whether the Steelers kick or go for two stands up to a much lower success probability. How low do people think the true probability is? Even if it's 10%, Pittsburgh's win probability is still only 0.006 lower if they go for two than if they kick. I just added a sentence to the article to make this point clear.

Bill Krasker

 

[Morton Muffley]

I knew footballcommentary.com would "comment" on this. Looks like, statistically, it's about the same whether you go for it or kick the XP.

A more interesting situation arose after the Steelers scored a touchdown with 10:25 left in regulation to pull within five points prior to the try. The two-point conversion apparrently succeeded, but Pittsburgh was flagged for holding, resulting in a re-try from the 12-yard line. The Steelers once again went for two, but failed.

According to the Chart, Pittsburgh should go for two in that situation if the probability of success exceeds about 0.19. That's close to the actual probability of scoring from the 12-yard line. So ostensibly, it's a close call. However, the Chart doesn't take into account that if the Steelers kick the extra point, it's a 30-yard attempt rather than the normal 20-yard attempt, and the success probability is only about 90% rather than 99%. This reduces the required probability of success on a two-point conversion from 0.19 to 0.17, and makes going for two slightly preferred despite the penalty. However, there is very little difference in win probability regardless of Pittsburgh's choice.

Call me stubborn, but I don't believe this can be correct. First, I find it hard to believe that the chances of success from the 12 yard line are anywhere near 19%. Where are they getting this data from? Surely, it isn't from a rich database of 2-point conversion from the twelve. So it must be some amalgam of other scenarios such as 3rd and 12 conversions. I think this is wrong because converting 3rd and twelve from midfield is certainly much less difficult than converting Tomlin's 2-point conversion because there is far less room to work with. That is, on 3rd and 12 from midfield you can complete a 30 yard pass. A 30 yard pass play was not an option for Pitt in their situation. Secondly, is a 30 yard FG/XP really only 90% likely or is this the likelihood of converting from 30-39 yards?

That is hardly a good argument in your favor.I dunno how exactly he calculated the odds of converting a 4th and goal from the 12, but I'd like to think he didn't just go by converting a 3rd and 12 from the 50 yard line. Most of the stuff on his site is over my head statistics wise, but I have to start with the assumption that the guy knows what he's talking about (and doing) and didn't just make up numbers for the heck of it.

Hey, KRS...from your most recent post it sounds like what I feared/suggested is pretty close to what he did. More importantly, thanks for getting and communicating this info. This is exactly what I love about this site.

And let me state clearly that I have no problem with someone taking a SWAG at the conversion success rate, but I do think he should have been clear about this from the outset. OK, so what have we got...well, he's provided almost EXACTLY what I assumed he provided...3rd down stats that could take place anywhere on the field. If we use 27% as the resonable number for 3 and 13 with the whole field (and 4th down avail...yes some will throw underneath in the hopes of breaking one or going for it on 4th and short), then I think 10% is a more reasonable figure to use as the conversion success rate for Pitt's 2-point attempt. Remember this is a short field, against a tough defense, occurring at a key point in a playoff game (yes, I think that converting this in the 4th quarter of a playoff game would be considerably tougher than converting something similar in the 2nd quarter of a non-playoff game).

OK, so now using the 10% likelihood figure we discover that Pitt's win percentage drops by less than 1% according to Mr. Krasker. Here's my second issue that I take with all those who follow the chart. I believe there is abig difference between using the chart with 3 minutes left vs. using it with 10 minutes left. As such, I question whether the chart Mr. Krasker is referring to takes into account how much time is left in the game. If it doesn't then this is a significant weakness. If it does then I'd like to see how a nearly 1 point swing (10%x2=.2 VS. 90%x1=.9) amounts to lowering a team's winning percentage by less than 1%. By my SWAG, I'd say throwing away .7 points (that's what Tomlin did, remember) should amount to something closer to a 10% reduction in winning percentage. Again, I've got nothing but gut to base that on so I'm probably way off base, but would love to see the math that shows how he came up with the .006 figure. Is that on their site? I didn't see it and am not asking you KRS to be the go-between, but if you or anyone else knows where this is great, let's take a look. Otherwise, I'll be happy to contact him myself.

Later.

 

[KnowledgeReignsSupreme]

His entire model is based on calculating your chance of winning in certain decisions (going for 2, going for it on 4th down) using time remaining, field position and score.

For instance, here is a go for it on 4th down table for the 2nd half if you are on the opponent's 40 yard line.

And here is the 2pt conversion chart. You'll see the minimum probability of success to go for it is determined by time and score differential.

At any rate, I think we can both agree that maybe (assuming only a 10% probability of success), costing your team .6% chance of winning the game is not the stupidest coaching decision ever.

 

[teamroc]

His entire model is based on calculating your chance of winning in certain decisions (going for 2, going for it on 4th down) using time remaining, field position and score.

For instance, here is a go for it on 4th down table for the 2nd half if you are on the opponent's 40 yard line.

And here is the 2pt conversion chart. You'll see the minimum probability of success to go for it is determined by time and score differential.

At any rate, I think we can both agree that maybe (assuming only a 10% probability of success), costing your team .6% chance of winning the game is not the stupidest coaching decision ever.

10% probability of success.6% chance of winning

Are someone's opinion. It's someone's theory. It's not fact. It's not the laws of math. Stop passing them off as such.

There is no equation to prove either side of the argument. It's silly to even try to come up with a magic formula.

Going for 2 from the 12 was a dumb idea. End of story.

 

[Da Guru]

His entire model is based on calculating your chance of winning in certain decisions (going for 2, going for it on 4th down) using time remaining, field position and score.

For instance, here is a go for it on 4th down table for the 2nd half if you are on the opponent's 40 yard line.

And here is the 2pt conversion chart. You'll see the minimum probability of success to go for it is determined by time and score differential.

At any rate, I think we can both agree that maybe (assuming only a 10% probability of success), costing your team .6% chance of winning the game is not the stupidest coaching decision ever.

10% probability of success.6% chance of winning

Are someone's opinion. It's someone's theory. It's not fact. It's not the laws of math. Stop passing them off as such.

There is no equation to prove either side of the argument. It's silly to even try to come up with a magic formula.

Going for 2 from the 12 was a dumb idea. End of story.

Nothing more needs to be said!

 

[menobrown]

The worst thing about him going for two is it seems to obscured what his real mistake was which was going too conservative on offense with a few minutes to go and the Jags having all timeouts left. I hate that no one in the media is questioning him about this and instead is focusing on the two point attempt. I like Tomlin but I want him to get grilled over this horrible mistake so maybe he'll be less inclined to make it again.

 

[Morton Muffley]

The worst thing about him going for two is it seems to obscured what his real mistake was which was going too conservative on offense with a few minutes to go and the Jags having all timeouts left. I hate that no one in the media is questioning him about this and instead is focusing on the two point attempt. I like Tomlin but I want him to get grilled over this horrible mistake so maybe he'll be less inclined to make it again.

Totally agree with this and said as much earlied in this thread. His decision to call the QB roll-out with no passing option greatly overestimated the value of forcing Jax to use a TO.

 

[Franknbeans]

The worst thing about him going for two is it seems to obscured what his real mistake was which was going too conservative on offense with a few minutes to go and the Jags having all timeouts left. I hate that no one in the media is questioning him about this and instead is focusing on the two point attempt. I like Tomlin but I want him to get grilled over this horrible mistake so maybe he'll be less inclined to make it again.

Totally agree with this and said as much earlied in this thread. His decision to call the QB roll-out with no passing option greatly overestimated the value of forcing Jax to use a TO.

Unfortunately, that was the OC's play call but obviously, Tomlin had the final say.

 

[Jiggyonthehut]

at people criticising the math when they can't come up with their own "intiuitive" models.

I don't this its unreasonable to say that, for example, a 4th and goal from the 12 yard line would be converted 15 times out of 100. I think the math yields a reasonable result.

 

[Morton Muffley]

His entire model is based on calculating your chance of winning in certain decisions (going for 2, going for it on 4th down) using time remaining, field position and score.

For instance, here is a go for it on 4th down table for the 2nd half if you are on the opponent's 40 yard line.

And here is the 2pt conversion chart. You'll see the minimum probability of success to go for it is determined by time and score differential.

At any rate, I think we can both agree that maybe (assuming only a 10% probability of success), costing your team .6% chance of winning the game is not the stupidest coaching decision ever.

Yeah, I saw thos charts after I posted the above and while I do believe that these charts are superior to what I imagine coaches like Tomlin are using. Call me an ignorant Pats fan, bu I have no doubt Belichik has his own stats guys that have created something similar (Ernie Adams perhaps?), but am pretty sure that many other coaches are using a much simpler one that does not distinguish between time remaining in the game. Perhaps I'm naive, but I have a hard time believing that coaches like Mike Tice and Dennis Green were this sophisticated.As for the charts themselves....they seem too aggressive to me. For example, they indicate that a team trailing by 10 with 12 minutes left in the game should go for 2 after scoring a TD. I'm not sure I've ever seen a coach do this because on its face it just seems very very wrong. I'd bet that the underlying stats are likely suggesting that what you gain by avoiding the 50/50 proposition that is OT is more than made up for by the egg on the coaches face the several times he ends up looking like a jackass because his is must go for a game-winning TD (down by four) instead of a game-tying FG (down by 3).

But the most nagging concern I have with this analysis is that it seems to be based on static success stats. What I mean is that in my hunt for these charts and the analysis behind them it seems that they are based on dynamic programming models which set the likely TD/FG chance of you an your opponent at a static figure. This seems to be blatantly wrong. For instance, these charts attempt to calculate the "right" decision based on the probability of future events (TD/FG/other) of the team and its opponent. Unfortunately, it seems that the model fails to recognize that the likelihood of these events probably changes dramatically in this key latter part of the game. For example, the models seem to use the constant 21% TD chance and 12% FG chance for both teams in order to determine the optimal point-after strategy. Well, if one team is trailing by 4 with 2 minutes left I can guarantedd you that their chance of kicking a FG is close to zero. In addition, I'd also suggest that the opposing teams chances of scoring a TD are likely much higher than 21% - mainly because I know the team will have an extra down to work with. Conversely, a team that is trying to milk the clock would expected to be somewhat lower than the 21%/12% likelihoods on which the models are built.

Now, I have no problem believing that teams should go for two more than they currently do...I just don't believe they should do it as often as these charts indicate (for the reasons cited above).

Anyway, it's a fascinating topic - at least to me. And I appreciate you bringing this analysis to my/our attention KRS.

As always, just my opinion.

 

[Morton Muffley]

at people criticising the math when they can't come up with their own "intiuitive" models. I don't this its unreasonable to say that, for example, a 4th and goal from the 12 yard line would be converted 15 times out of 100. I think the math yields a reasonable result.

I'm not sure I understand your point. Can you please clarify?

 

[Morton Muffley]

The worst thing about him going for two is it seems to obscured what his real mistake was which was going too conservative on offense with a few minutes to go and the Jags having all timeouts left. I hate that no one in the media is questioning him about this and instead is focusing on the two point attempt. I like Tomlin but I want him to get grilled over this horrible mistake so maybe he'll be less inclined to make it again.

Totally agree with this and said as much earlied in this thread. His decision to call the QB roll-out with no passing option greatly overestimated the value of forcing Jax to use a TO.

Unfortunately, that was the OC's play call but obviously, Tomlin had the final say.

Of course we realize it was the OC play call, but we also realize that Tomlin factored heavily in defining the type of play that was called (i.e. very aggressive, somewhat aggresive, somewhat conservative, very conservative). And that he rubber-stamped the actually play called.I think the argument that I'm supporting was the choice Tomlin made to go "very conservative" in this instance when he should have gone "somewhat aggressive".

 

[Topes]

Going for 2 with 10:25 to play was boneheaded. The distance doesn't matter. The chart doesn't matter.

No chart can account for how both teams are playing.

10:25 to go leaves the distinct possibility that your opponent will get a FG.

Options:

Kick the XP. Down by 4. If opp scores a FG, you've got a TD to go.

Try the 2P.

 

[Da Guru]

Going for 2 with 10:25 to play was boneheaded. The distance doesn't matter. The chart doesn't matter.No chart can account for how both teams are playing.10:25 to go leaves the distinct possibility that your opponent will get a FG.Options:Kick the XP. Down by 4. If opp scores a FG, you've got a TD to go.Try the 2P.

I agree...a chart can`t account for the pressure of the game, a missed block, a WR slipping, a fumbled snap, the defense blitzing 2..3..or 4 guys. A zone or a man to man. The field conditions.Way too many variables.From the 2..go for two. From the 12 with 10:30 to play and your defense playing well take the sure thing and kick the PAT.

 

[lukerguy]

Ill put my two cents in:

1) Would I have gone for 2 in the first place? no. I specifically said at the time to guys I was watching the game with that the move was a mistake and shouldnt be done.

However, and a very big however, it could have turned out well, and would have forced Jacksonville to score a TD, maybe the situations would have changed.

2) The biggest problem I saw with Tomlin's decision making was his 3 min strategy. I don't know why coach after coach tries to run the clock out. it's not really that hard to defend, put 8 or 9 in the box and stop the run. Pitt came back from how many in the 4th quarter doing what? THROWING the ball.

With 3 mins left Jacksonville had a full set of time outs, did he really think going 3 and out would hurt them. I mean as if 2 mins and a time out isn't enough for any team to go 80 yrds, let alone 25 for a FG.

Anyone ever realize why the patriots win playoff games? When they have the ball and the lead late in a close game they dont get conservative. Or another example, anyone watch the LSU vs Ohio St game? late in the game, when LSU was up 14 (right after OSU scored) they continued to go 4 WR sets and finally iced the game after a couple 1st downs and eventually a TD.

If Tomlin called the last 3 mins like the first 12 mins of the Q it would be Pitt vs NE on Sat.

 

[Godsbrother]

2) The biggest problem I saw with Tomlin's decision making was his 3 min strategy. I don't know why coach after coach tries to run the clock out. it's not really that hard to defend, put 8 or 9 in the box and stop the run. Pitt came back from how many in the 4th quarter doing what? THROWING the ball. With 3 mins left Jacksonville had a full set of time outs, did he really think going 3 and out would hurt them. I mean as if 2 mins and a time out isn't enough for any team to go 80 yrds, let alone 25 for a FG. Anyone ever realize why the patriots win playoff games? When they have the ball and the lead late in a close game they dont get conservative. Or another example, anyone watch the LSU vs Ohio St game? late in the game, when LSU was up 14 (right after OSU scored) they continued to go 4 WR sets and finally iced the game after a couple 1st downs and eventually a TD.If Tomlin called the last 3 mins like the first 12 mins of the Q it would be Pitt vs NE on Sat.

I agree 100%. That is where the game as lost. It was especially disappointing considering how well the Steelers had been moving the ball through the air. Why not give yourself at least one chance to get the 1st down and put the game away?What made it worse was as bad as the Steeler special teams were you knew Jax was going to great field position if you went 3-and-out. Sure enough, Jax started their drive on the 49 yard line. Pitiful...

 

[KnowledgeReignsSupreme]

His entire model is based on calculating your chance of winning in certain decisions (going for 2, going for it on 4th down) using time remaining, field position and score.

For instance, here is a go for it on 4th down table for the 2nd half if you are on the opponent's 40 yard line.

And here is the 2pt conversion chart. You'll see the minimum probability of success to go for it is determined by time and score differential.

At any rate, I think we can both agree that maybe (assuming only a 10% probability of success), costing your team .6% chance of winning the game is not the stupidest coaching decision ever.

10% probability of success.6% chance of winning

Are someone's opinion. It's someone's theory. It's not fact. It's not the laws of math. Stop passing them off as such.

There is no equation to prove either side of the argument. It's silly to even try to come up with a magic formula.

Going for 2 from the 12 was a dumb idea. End of story.

You might want to switch to decaf.

 

[fred_1_15301]

2) The biggest problem I saw with Tomlin's decision making was his 3 min strategy. I don't know why coach after coach tries to run the clock out. it's not really that hard to defend, put 8 or 9 in the box and stop the run. Pitt came back from how many in the 4th quarter doing what? THROWING the ball. With 3 mins left Jacksonville had a full set of time outs, did he really think going 3 and out would hurt them. I mean as if 2 mins and a time out isn't enough for any team to go 80 yrds, let alone 25 for a FG. Anyone ever realize why the patriots win playoff games? When they have the ball and the lead late in a close game they dont get conservative. Or another example, anyone watch the LSU vs Ohio St game? late in the game, when LSU was up 14 (right after OSU scored) they continued to go 4 WR sets and finally iced the game after a couple 1st downs and eventually a TD.If Tomlin called the last 3 mins like the first 12 mins of the Q it would be Pitt vs NE on Sat.

I agree 100%. That is where the game as lost. It was especially disappointing considering how well the Steelers had been moving the ball through the air. Why not give yourself at least one chance to get the 1st down and put the game away?What made it worse was as bad as the Steeler special teams were you knew Jax was going to great field position if you went 3-and-out. Sure enough, Jax started their drive on the 49 yard line. Pitiful...

godsbrother....I'm not living in Pittsburgh anymore. What was the reaction to Tomlin's conservative coaching on that drive by the media, sports radio hosts (i.e. Mark Madden) and the fans? Did most people criticize his decisions on that drive?

 

[KnowledgeReignsSupreme]

Yeah, I saw thos charts after I posted the above and while I do believe that these charts are superior to what I imagine coaches like Tomlin are using. Call me an ignorant Pats fan, bu I have no doubt Belichik has his own stats guys that have created something similar (Ernie Adams perhaps?), but am pretty sure that many other coaches are using a much simpler one that does not distinguish between time remaining in the game. Perhaps I'm naive, but I have a hard time believing that coaches like Mike Tice and Dennis Green were this sophisticated.

I agree here. I doubt there's more than 5 or 6 coaches in the whole league who do this. Belichick and Reid strike me as two that would do this. Dungy also.

As for the charts themselves....they seem too aggressive to me. For example, they indicate that a team trailing by 10 with 12 minutes left in the game should go for 2 after scoring a TD. I'm not sure I've ever seen a coach do this because on its face it just seems very very wrong. I'd bet that the underlying stats are likely suggesting that what you gain by avoiding the 50/50 proposition that is OT is more than made up for by the egg on the coaches face the several times he ends up looking like a jackass because his is must go for a game-winning TD (down by four) instead of a game-tying FG (down by 3).

You're reading the chart wrong. The -10 means the score differential AFTER the TD is scored. It means they were down by 16 with 12 minutes left before the TD. The scenario you're talking about would be the -4 row, which has a much higher chance of success to go for it rate.But let's touch on the subject of "too aggressive" in general You can't use the fact that coaches are afraid to do this against the model necessarily. Coaches are concerned with keeping their job first and foremost. That means sticking with tradition and not being too aggressive. The aggressive playcall that fails, even if it was a +EV playcall, is criticized. The conservative, this is what all coaches do, -EV playcall that fails is not criticized.

Let's look at more specific "aggressive" example. Down 15 (after the TD) with 30 minutes to go (assume a kick return TD), the chart says you need a .41 chance of success. 2pt conversions are at (according to him) 43% right now. How many coaches do you see going for 2 to start the 2nd half? And lots of fans and commentators will say "no it's too early to go for 2". Going for 2 now is "too aggressive". But if you think about it, how many teams down 2 TDs at the start of the half go on to win anyway? It's like the coaches who punt down 3 TDs in the 2nd half. Why? This isn't college football. Keeping the score close in a loss doesn't improve your poll ranking. If you fail, so what? Go for it. Scare their D a little. Make them stay on the field a little more so they are tired next series.

If you look at successful coaches like Reid and Belichick, you'll notice they are much more "aggressive" in situations like this. They go for it on 4th down more, etc etc. Having good job security allows them to make the right calls more often. "Of course they should've gone for it on 4th down there, Belichick is a genius". Just like Gruden going for 2 against the Redskins a couple years ago, and Vermeil going for 2 against the Raiders that same year (which coincidently is the exact same situation that the Ravens messed up this year).

But the most nagging concern I have with this analysis is that it seems to be based on static success stats. What I mean is that in my hunt for these charts and the analysis behind them it seems that they are based on dynamic programming models which set the likely TD/FG chance of you an your opponent at a static figure. This seems to be blatantly wrong. For instance, these charts attempt to calculate the "right" decision based on the probability of future events (TD/FG/other) of the team and its opponent. Unfortunately, it seems that the model fails to recognize that the likelihood of these events probably changes dramatically in this key latter part of the game. For example, the models seem to use the constant 21% TD chance and 12% FG chance for both teams in order to determine the optimal point-after strategy. Well, if one team is trailing by 4 with 2 minutes left I can guarantedd you that their chance of kicking a FG is close to zero. In addition, I'd also suggest that the opposing teams chances of scoring a TD are likely much higher than 21% - mainly because I know the team will have an extra down to work with. Conversely, a team that is trying to milk the clock would expected to be somewhat lower than the 21%/12% likelihoods on which the models are built.

I have no idea if this is true for FC. You'll have to e-mail him and ask.

 

[Godsbrother]

2) The biggest problem I saw with Tomlin's decision making was his 3 min strategy. I don't know why coach after coach tries to run the clock out. it's not really that hard to defend, put 8 or 9 in the box and stop the run. Pitt came back from how many in the 4th quarter doing what? THROWING the ball. With 3 mins left Jacksonville had a full set of time outs, did he really think going 3 and out would hurt them. I mean as if 2 mins and a time out isn't enough for any team to go 80 yrds, let alone 25 for a FG. Anyone ever realize why the patriots win playoff games? When they have the ball and the lead late in a close game they dont get conservative. Or another example, anyone watch the LSU vs Ohio St game? late in the game, when LSU was up 14 (right after OSU scored) they continued to go 4 WR sets and finally iced the game after a couple 1st downs and eventually a TD.If Tomlin called the last 3 mins like the first 12 mins of the Q it would be Pitt vs NE on Sat.

I agree 100%. That is where the game as lost. It was especially disappointing considering how well the Steelers had been moving the ball through the air. Why not give yourself at least one chance to get the 1st down and put the game away?What made it worse was as bad as the Steeler special teams were you knew Jax was going to great field position if you went 3-and-out. Sure enough, Jax started their drive on the 49 yard line. Pitiful...

godsbrother....I'm not living in Pittsburgh anymore. What was the reaction to Tomlin's conservative coaching on that drive by the media, sports radio hosts (i.e. Mark Madden) and the fans? Did most people criticize his decisions on that drive?

Madden definitely did and so did Stan Savran. I didn't hear Junker and Crowe's reaction but I assume they were critical as well.

 

[Morton Muffley]

You're reading the chart wrong. The -10 means the score differential AFTER the TD is scored. It means they were down by 16 with 12 minutes left before the TD. The scenario you're talking about would be the -4 row, which has a much higher chance of success to go for it rate.But let's touch on the subject of "too aggressive" in general You can't use the fact that coaches are afraid to do this against the model necessarily. Coaches are concerned with keeping their job first and foremost. That means sticking with tradition and not being too aggressive. The aggressive playcall that fails, even if it was a +EV playcall, is criticized. The conservative, this is what all coaches do, -EV playcall that fails is not criticized.Let's look at more specific "aggressive" example. Down 15 (after the TD) with 30 minutes to go (assume a kick return TD), the chart says you need a .41 chance of success. 2pt conversions are at (according to him) 43% right now. How many coaches do you see going for 2 to start the 2nd half? And lots of fans and commentators will say "no it's too early to go for 2". Going for 2 now is "too aggressive". But if you think about it, how many teams down 2 TDs at the start of the half go on to win anyway? It's like the coaches who punt down 3 TDs in the 2nd half. Why? This isn't college football. Keeping the score close in a loss doesn't improve your poll ranking. If you fail, so what? Go for it. Scare their D a little. Make them stay on the field a little more so they are tired next series.If you look at successful coaches like Reid and Belichick, you'll notice they are much more "aggressive" in situations like this. They go for it on 4th down more, etc etc. Having good job security allows them to make the right calls more often. "Of course they should've gone for it on 4th down there, Belichick is a genius". Just like Gruden going for 2 against the Redskins a couple years ago, and Vermeil going for 2 against the Raiders that same year (which coincidently is the exact same situation that the Ravens messed up this year).

Thanks KRS, I WAS reading the chart wrong. Good thing I wasn't in the heat of a playoff game, huh? The worst part is that I did this the first time I read the chart and then realized my mistake...then went ahead and posted my originally mistake here anyway. Ugh! Anyway, the one that I wanted to point out is that the chart states that early in the 4th quarter when down by 14 the trailing team should go for two after scoring a TD. That is, when you are down eight you should go for two rather than try to cut the deficit to seven. I've run the numbers and they make sense if you assume that you've got a 50/50 chance in OT. So the aggressive strategy of going for 2 in this situation requires a willingness for the coach to lessen his chances of forcing overtime in order to increase in his chances of winning the game outright. The math comes out that going for two gives you a 2% better chance of winning the game outright. Again this all assumes that everything is independent of one another and that there is no such thing as momentum (as well as some other assumptions that I'm willing to accept). So you are absolutely corect that it comes down to how much second-guessing does the coach want to take when his strategy fails (as it will often) and even when it succeeds. Has anyone seen a coach for two when trailing by eight rather than kicking the XP to make it a 7 point game? I can't recall this.lastly, as a Pats fan who watches alot of NFL football I'm continually amazed by the fact that other coaches haven't picked up on the aggressiveness that the Pats show on a weekly basis. Here's a clue for those coaches...the Pats aren't aggressive because they are good...the Pats are good because they are aggressive. Well, technically, the Pats are great this year...but for all those years that that had average talent and squeaked out win after win, it's because they didn't do stupid things like punt on 4th and 2 from the opponents 40 yard line. I still can't believe teams continue to do this.

 

[Jiggyonthehut]

at people criticising the math when they can't come up with their own "intiuitive" models. I don't this its unreasonable to say that, for example, a 4th and goal from the 12 yard line would be converted 15 times out of 100. I think the math yields a reasonable result.

I'm not sure I understand your point. Can you please clarify?

Yes. One poster above used a 3% chance of conversion guesstimate as referenced by a Pitt beat writer. Now that number is completely absurd. Think of the 2-point attempt as an isolated 4th-goal play from the 12 yard line. Assuming the team is attempting to score a touchdown and not turn the ball over, that is. They've got one shot only. A 3% probability infers that the team would only be successful 3 attempts out of 100. I hope people can understand that this is pretty much out of the realm of possibility. If it were to occur, it would be likely that the person calling the plays will likely never get a job on any level of organized football. Converting 15 times out of 100 is equivalent to about a 1 in 7 change of scoring the touchdown. This seems, to me, to be pretty attainable. I feel that if an OC were to go 0 for 7 in such situations, his job probably would be on the hot seat.

 

[Morton Muffley]

at people criticising the math when they can't come up with their own "intiuitive" models. I don't this its unreasonable to say that, for example, a 4th and goal from the 12 yard line would be converted 15 times out of 100. I think the math yields a reasonable result.

I'm not sure I understand your point. Can you please clarify?

Yes. One poster above used a 3% chance of conversion guesstimate as referenced by a Pitt beat writer. Now that number is completely absurd. Think of the 2-point attempt as an isolated 4th-goal play from the 12 yard line. Assuming the team is attempting to score a touchdown and not turn the ball over, that is. They've got one shot only. A 3% probability infers that the team would only be successful 3 attempts out of 100. I hope people can understand that this is pretty much out of the realm of possibility. If it were to occur, it would be likely that the person calling the plays will likely never get a job on any level of organized football. Converting 15 times out of 100 is equivalent to about a 1 in 7 change of scoring the touchdown. This seems, to me, to be pretty attainable. I feel that if an OC were to go 0 for 7 in such situations, his job probably would be on the hot seat.

Agree that 3% is crazy low. I think football commentary was estimating it to be around 19% which seemed to high to me and caused me to question where they got their stats from. One in seven seems a bit high to me as I thought it might be closer to one in ten, but wouldn't be surprised if it were one in seven. Anyway, thanks for the clarification.

 

[Steeler]

2) The biggest problem I saw with Tomlin's decision making was his 3 min strategy. I don't know why coach after coach tries to run the clock out. it's not really that hard to defend, put 8 or 9 in the box and stop the run. Pitt came back from how many in the 4th quarter doing what? THROWING the ball.

With 3 mins left Jacksonville had a full set of time outs, did he really think going 3 and out would hurt them. I mean as if 2 mins and a time out isn't enough for any team to go 80 yrds, let alone 25 for a FG.

Anyone ever realize why the patriots win playoff games? When they have the ball and the lead late in a close game they dont get conservative. Or another example, anyone watch the LSU vs Ohio St game? late in the game, when LSU was up 14 (right after OSU scored) they continued to go 4 WR sets and finally iced the game after a couple 1st downs and eventually a TD.

If Tomlin called the last 3 mins like the first 12 mins of the Q it would be Pitt vs NE on Sat.

I agree 100%. That is where the game as lost. It was especially disappointing considering how well the Steelers had been moving the ball through the air. Why not give yourself at least one chance to get the 1st down and put the game away?What made it worse was as bad as the Steeler special teams were you knew Jax was going to great field position if you went 3-and-out. Sure enough, Jax started their drive on the 49 yard line. Pitiful...

godsbrother....I'm not living in Pittsburgh anymore. What was the reaction to Tomlin's conservative coaching on that drive by the media, sports radio hosts (i.e. Mark Madden) and the fans? Did most people criticize his decisions on that drive?

You can listen to ESPN Radio... click the "Listen Live" link... the player that pops up has a Pittsburgh tab... you can hear the Gunker & Crowe, and the Mark Madden shows (maybe others, but I only listen while at work).

 

[pecorino]

everybody and their brother here in pittsburgh has been lambasting the coaching staff all week. the complaints usually focus more on the poor playing calling down the stretch than the two-point conversion from the 12.

earlier i mentioned the media guy who threw out the 3% number. i agree that it is too low, but my own guess is that he is closer than the folks who had it at 19%. i'd say 10% myself but with no data to go on, it is a crap shoot.

 

[Fiddles]

im confident tomlin learned from this and it wont happen again

 

[Maurile Tremblay]

In his post-game press conference, airing on ESPN right now, he said he went for 2 then because that is what the chart says and also said that everyone has the same chart. Um, okay.

He also said if he had a crystal ball and knew they would lose by two he would have kicked the XP if that makes you feel better.

Situations change every possesion of the game. Tomlin made a blunder. Regardless of the final score.

Exactly. Which is why his crytal ball argument is so insulting, because it assumes that there weren't football fans all across the country screaming/questioning/wondering why he would continue to go for two after being backed up ten yards. Does his chart tell him to go for two if he's backed up 20 yards? what about 50 yards? 80 yards? see this is the problem...the chart ASSUMES the all is on the 2 1/2 yard line.

I don't know about 20 yards, but if he's backed up 50 yards or especially 80 yards, he should almost certainly go for two (regardless of the score or time remaining).

 

[Maurile Tremblay]

As best I can tell, there have been 20 fourth-and-goal attempts from between 10 and 15 yards out between 2002 and 2007. Six out of those 20 plays resulted in touchdowns. So on a very limited sample, that's 30 percent. (Data dominator query.)

 

[Sheriff66]

2) The biggest problem I saw with Tomlin's decision making was his 3 min strategy. I don't know why coach after coach tries to run the clock out. it's not really that hard to defend, put 8 or 9 in the box and stop the run. Pitt came back from how many in the 4th quarter doing what? THROWING the ball. With 3 mins left Jacksonville had a full set of time outs, did he really think going 3 and out would hurt them. I mean as if 2 mins and a time out isn't enough for any team to go 80 yrds, let alone 25 for a FG. Anyone ever realize why the patriots win playoff games? When they have the ball and the lead late in a close game they dont get conservative. Or another example, anyone watch the LSU vs Ohio St game? late in the game, when LSU was up 14 (right after OSU scored) they continued to go 4 WR sets and finally iced the game after a couple 1st downs and eventually a TD.If Tomlin called the last 3 mins like the first 12 mins of the Q it would be Pitt vs NE on Sat.

I agree 100%. That is where the game as lost. It was especially disappointing considering how well the Steelers had been moving the ball through the air. Why not give yourself at least one chance to get the 1st down and put the game away?What made it worse was as bad as the Steeler special teams were you knew Jax was going to great field position if you went 3-and-out. Sure enough, Jax started their drive on the 49 yard line. Pitiful...

godsbrother....I'm not living in Pittsburgh anymore. What was the reaction to Tomlin's conservative coaching on that drive by the media, sports radio hosts (i.e. Mark Madden) and the fans? Did most people criticize his decisions on that drive?

People are calling it what it was, no balls playcalling...The Turtle Offense.

 

[BLOX]

I was watching the replay on NFLN the other night and realized something about the 2-pt try from the 12.

It is an untimed down that a turnover cannot be returned for a score on. Basically, what I don't understand is why Roethlisberger wasn't aware of this. He was turned around at the 3 yard line. He could have pitched the ball back to someone. It wouldn't matter if the defense recovered the ball as he was stopped anyway. However, since the Def can't advance the ball after recovering it the offense should try to keep the ball alive until someone has a path to the end zone. No one, however has ever done this so I can't say it was bad coaching, but it's interesting that a game of keep away doesn't start once a RB or WR is stopped short of the end zone.

 

[Sheriff66]

Tomlin says he would pass on the 3rd down play, but go for two again.

Had they the chance to do it again, Tomlin said the Steelers probably would try a pass on third-and-6 on their next-to-last drive against Jacksonville when they had the lead and the ball. Instead, Ben Roethlisberger rolled right and was tackled after a 1-yard gain.

"In hindsight, if you had it to do over, yes you would love to do that," Tomlin said, "because the result of the play that we choose was not the result we were looking for. We weren't successful."

On the other hand, he said he would still try the two-point conversion even with the ball perched on the 12-yard line with 10:29 to play.

"Yes, I'd do it again. I throw caution to the wind. I play to win."

First published on January 11, 2008 at 12:00 am

 

[KnowledgeReignsSupreme]

In his post-game press conference, airing on ESPN right now, he said he went for 2 then because that is what the chart says and also said that everyone has the same chart. Um, okay.

He also said if he had a crystal ball and knew they would lose by two he would have kicked the XP if that makes you feel better.

Situations change every possesion of the game. Tomlin made a blunder. Regardless of the final score.

Exactly. Which is why his crytal ball argument is so insulting, because it assumes that there weren't football fans all across the country screaming/questioning/wondering why he would continue to go for two after being backed up ten yards. Does his chart tell him to go for two if he's backed up 20 yards? what about 50 yards? 80 yards? see this is the problem...the chart ASSUMES the all is on the 2 1/2 yard line.

I don't know about 20 yards, but if he's backed up 50 yards or especially 80 yards, he should almost certainly go for two (regardless of the score or time remaining).

I missed his poor logic the first time on that. Thanks.

 

[KnowledgeReignsSupreme]

I was watching the replay on NFLN the other night and realized something about the 2-pt try from the 12.It is an untimed down that a turnover cannot be returned for a score on. Basically, what I don't understand is why Roethlisberger wasn't aware of this. He was turned around at the 3 yard line. He could have pitched the ball back to someone. It wouldn't matter if the defense recovered the ball as he was stopped anyway. However, since the Def can't advance the ball after recovering it the offense should try to keep the ball alive until someone has a path to the end zone. No one, however has ever done this so I can't say it was bad coaching, but it's interesting that a game of keep away doesn't start once a RB or WR is stopped short of the end zone.

Heat of battle and all. Probably escaped the coach too. Great point.

 

[lukerguy]

Tomlin says he would pass on the 3rd down play, but go for two again.

Had they the chance to do it again, Tomlin said the Steelers probably would try a pass on third-and-6 on their next-to-last drive against Jacksonville when they had the lead and the ball. Instead, Ben Roethlisberger rolled right and was tackled after a 1-yard gain.

On the other hand, he said he would still try the two-point conversion even with the ball perched on the 12-yard line with 10:29 to play.

"Yes, I'd do it again. I throw caution to the wind. I play to win."

That's what is puzzling about the whole situation. A coach that plays to win doesnt try and ice the game with 3 runs in a row, that's playing not to lose. What's crazy to me about this is the amount of time left. Lets say they convert on 3rd and 6, if they went 3 and out after that jacksonville still would have got the ball back with 1 min + left. I would personally rather go 3 and out throwing the ball and waste none of the 3 mins of the clock then run it 3 times and make them use 2 of their time outs and still have a full 2 mins. All he did was hurt his chances of coming back and almost ensuring jacksonville with the last drive.

 

[KnowledgeReignsSupreme]

jiggyonthehut doing good work in here.