XFL Point After Way Too Early Analytics
I've been intrigued by the XFL's use of 1pt from the 2yd, 2pt from the 5yd, and 3pt from the 10yd. So I poured over 8 box scores worth of data to put together some information.
Conclusion so far: going for 2pt from the 5yd is worth almost half a point more to XFL teams than going for 1pt from the 2yd.
*So far only two 3pt from the 10yd attempts have been made, and on one the defense had a too many men penalty that had 5 yards enforced so the scoring play was from the 5 so I decided to count that as a 2pt attempt.
In summary:
For 1pt attempts the league is 6 makes out of 23 attempts. For an expected value of 0.261 points per attempt.
For 2pt attempts the league is 5 makes out of 14 attempts. For an expected value of 0.714 points per attempt.
For 3pt attempts the leagues is 0 for 1.
It will be interesting to see if XFL teams start to go for 2pts more often as the season progresses.
February 18th, 2020 at 12:54 PM ^
From a broad standpoint, it looks like most of these coaches are those who are old-fashioned and have been around for a long time (minus Pep Hamilton).
I wouldn't be surprised if they ignore analytics and continue to go for 1 even though going for 2 clearly seems like the better option.
In a weird way, going from the 5 yard line almost makes it easier for the offense because there's more room
February 18th, 2020 at 1:48 PM ^
That sample size is far too small to conclude that going for 2 is clearly the better option.
February 18th, 2020 at 2:39 PM ^
The sample size is, but basic logic and estimation would bare it out to be the obviously superior choice.
Let's assume teams would convert the 2 pointer 33% of the time (which is a conservative estimate imo. That's .667 points per attempt.
For the 1 pointer to be superior - teams would have to convert more than 67% of the time.
February 18th, 2020 at 6:58 PM ^
If the league survives long enough we may get a large enough sample size?
But I think if 2 pt Xtras are that much more advantageous they may consider making the 1pt try from the 1.
February 19th, 2020 at 8:24 AM ^
My guess is that there are a variety of factors in play here. First and foremost is the play-calling from the different distances. My guess is that if the same plays were being called from the 2 as from the 5 the one-point conversions would be much higher.
February 18th, 2020 at 3:17 PM ^
As KT noted, it's about the math behind it.
February 18th, 2020 at 3:10 PM ^
Pep Hamilton has shown terrible decision making in the extra point.
In the first week they went for 1 AFTER a defensive penalty was enforced half the distance to the goal. So then the options were 3 points from the 5, 2 points from the 2.5, or 1 point from the 1.
Pep Hamilton chose to go for 1.
February 18th, 2020 at 3:40 PM ^
DC is 1/5 on 1pt attempts (including the attempt from the 1yd line), 0/1 on 2pt attempts, and 0/1 on 3pt attempts. Not good.
February 18th, 2020 at 4:58 PM ^
What was the score and game situation?
That's a factor that's not discussed here, but it matters. When a team goes up by three in the fourth quarter, it doesn't matter if the expected points are greater for a 2-point conversion than a 1. That extra point, making a field goal useless, is so valuable that any increase in probability to gain it may be worth it in such cases.
I haven't watched much XFL, but I saw this over the weekend, when in a 6-3 game game the trailing team scored a TD in the fourth quarter and went for (and failed to gain) a single, followed by the other team doing the same. They did this despite knowing that the conversion would be a challenge, because it's still easier to score from closer in, and the value of that single point is immense.
I think we need to wait for firmer numbers before drawing any strong conclusions.
February 18th, 2020 at 3:21 PM ^
I'd be curious to see what Trestman does. He spent a long time in the CFL and managed to win three Grey Cups. I'd think that would make him more open to trying out some of the new features in XFL play.
February 18th, 2020 at 12:57 PM ^
Not sure that I agree with your methodology on counting the 3PA as a 2PA due to the penalty. I would think that part of the calculus on electing a 3-point play would be the marginal benefit of a defensive penalty. If the defense takes a holding or PI call, you're now going for 3 from the 1 yard line, or in the case you referenced, from the 5 in the case of too many men, offsides, etc.
February 18th, 2020 at 1:50 PM ^
I just figured the play was from 5 yards, so even though it was for 3pts it was from the 2pt yardage and better categorized with a 2pt attempts to show success from the 5yd line. If you take that one attempt out the 2pt attempts go from 5/14 EV of 0.714 to 4/13 EV 0.615. Still over a third of a point more in expected value for going for 2.
The sample sizes are still way to small. But for now it just seemed way to strange to say that 3pt attempts were 1/2 when the successful one's scoring play was from the 5 yard line after a penalty. I have a note of it in my spreadsheet.
February 18th, 2020 at 6:36 PM ^
With the fix, EV of the three point try is 1.5. Small sample size though.
February 18th, 2020 at 1:10 PM ^
don't you have anything better with your time? Maybe solving world peace or something.
February 18th, 2020 at 1:58 PM ^
What if XFL analytics lead to world peace?
February 18th, 2020 at 2:02 PM ^
Thumbs up for the complete lack of self awareness.
February 18th, 2020 at 1:19 PM ^
How are teams only scoring 25% of the time from the 2 yard line? Don't CFB and NFL teams average about 50% on 2 pt conversions?
February 18th, 2020 at 1:24 PM ^
Probably mostly sample size. I’d expect that to rise over time.
February 18th, 2020 at 1:25 PM ^
Always go for three you cowards!
February 18th, 2020 at 1:42 PM ^
Just going 1/4 from the 10 yard line is still better than converting at the league average from the 5 yard line.
February 18th, 2020 at 2:43 PM ^
LOL at the idea that a team in the XFL can score 4 touchdowns in one game. In 16 tries, how many times has it been done?
February 18th, 2020 at 1:50 PM ^
It's an interesting question. I wonder if you could get a better sample size if you used college or NFL data for 4th down and goal conversions from the 2, 5, and 10 yard lines. You would have to think someone did that analysis for the XFL when they went about picking those yard lines, but maybe not.
February 18th, 2020 at 2:58 PM ^
You'd assume that they tried to equalize the expected value if that were the case. There are 35 more games. It will be interesting to see how a full season of numbers adds up.
February 18th, 2020 at 4:16 PM ^
The 3 point conversion is due and Never Tell Me the Odds!