Wait, the analytics guy is saying take the points? Almost. *Keep* the points. The Raiders made a 3rd quarter field goal to go up 6 on 4th and 7.5. The Chiefs had a defensive penalty and Las Vegas elected to take the points off and go for it on... 4th and 2.5. (exact distance from Next Gen Stats). They failed and ended up losing by 3. When comparing the options of a made field goal vs a go for it attempt with 2.5 yards to go (the extra half yard matters), I have three public models and one non-public model all saying that keeping the points on the board was preferred from a win probability standpoint by a small margin. It is quite rare a team is "too aggressive", but I would say there is data supporting that is true here.show more

Ya a lot of 4th down decision making is centered around FGs not being an 100% success rate like many make it seem. However when it is 100% (keeping the points) make sense that it would give keeping the points an edge

Yes a different decision

Very cool stuff. Is your model public?

Curious, does the data factor in who the opponent is? Like, is there an adjustment for the "you have to get TDs and can't settle for FGs against Mahomes" thought process?

Does this take into account that specific team's ability to make short yardage? Feel like sometimes do it easily while others really suck at it

In this moment I am anchoring off of a league average for 2.5 yards to go. It is really hard to factor in how good a team is on offense with that exact distance, how good a defense is in that exact distance, those play callers, those QBs, etc.

What about kicking the 48 yarder on 4th and 1 later in the game

Would have to look at the moment

I felt sick when we missed it on 4th down after this Gut punch

Possibility of missing the field goal is a not-inconsequential part of the go-for-it math

Does your model factor in the live Vegas spread (or realistic chance to still win) at that moment? An important factor. If they were say still 10/1 underdogs (extreme example) they should be getting more aggressive even if it in a vacuum it says to keep the points.

Yes it does

If they had attempted and made FG, end of game would’ve played out differently had it just been a one-score game or had Chiefs had ball with tied game.

Is there a Mahomes factor in the models?

Your not going to beat the chiefs with field goals!!

Says who

@TheRaider509 Game would've been 30-30 raiders driving to get into field goal range. Changes picture dramatically.

@RyanPaganetti Maybe for some

How are these models made? And is there somewhere can I find these online?

Out of curiosity, what do the models say if its was 4th and 1 instead of 2.5?

Go for it

Someone should look at Dan Campbell's decision to go for a TD in the 4th Qtr down 13 with 7min left on 4th and 10 from the 15 (which I think was the right call) but then after TWO CONSECUTIVE 5 yd penalties he GOES FOR IT ANYWAY on 4th and 20!! vs. a FG

I will

Do you like epa or win probability to make these decisions?

The earlier in the game EPA is reasonable as it won’t be far from win probability but you ideally are making most of these off win probability, particularly second half

Just having fun with hypotheticals: suppose he missed the field goal (or play was blown dead before he took the kick). So now it’s 4th and 2.5. Would you attempt the FG or go for it? Thanks. 🙏

It is basically 50/50 there

When you're 3-0, haven't sniffed the playoffs in 10 years and you're playing your rival in the midst of a sb run all that goes out the window. Good call, coach.

When does it flip to go for it? 2 yards? 1.5? 1?

The losing by 3 isn’t an honest assessment. Chiefs went for 2 after the FG was taken off the board. If the FG stood they would have been up 4 not 3.

what if this was in the second quarter?

More inclined to go for it but still not clear to do it

Analytics won’t tell you that the Raiders took their own momentum away by taking points off the board,,,,,,

I’m less into the momentum aspect and more into the pure win probability

Oh I was joking lol