You have the O and D possession-based conversion percentages above. On a per-point basis, our O line had an 84% conversion, while D line had 78% (on points where they got a turnover). I think what this means is that the O line was better at getting the disc back after a turn, while the D line, if they got a chance and turned it, generally didn't get another one. This is borne out by another stat, which is that on points where the D line got a turn, they averaged 1.11 possessions. So it was pretty rare that they got more than one possession. That all makes sense to me since a multi-turn point becomes very tiring/taxing on the defense.
Obviously we went on a great run that tournament, and our O line's ability to get the disc back and D-lines ability to punch in breaks was a big part of it. Our efficiency in the bracket was also much higher overall.
I think a significant factor could be field position at the start of the possession. O-line possessions are likely to start from the pull around the brick or front of the end zone they are defending. Whereas defensive possessions are more likely to be further down the field after getting a turnover. Since the defence tends to have less distance to go before scoring, they are more likely to score before turning over the disk.
Yeah that's a good point. IDK if there's any USAU data that has field position, but UFA data could definitely be used to investigate that hypothesis.
I think the other big factor in my mind was that D-lines are playing offense against units that mostly practice playing offense (O-lines) while O-lines are playing offense against units that mostly practice playing D (D-lines).
So D-lines might not be stronger playing offense "in a vacuum", but they do sometimes have the relative advantage due to the frisbee meta of putting O-focused players on O-lines
This makes me wonder, how important is the starting position for when the defence starts their offensive possession? Would teams that get more D's on under & handler space throws have a higher conversion percentage than teams that get most of their D's on deep shots? This is similar to asking how effective is the "punt and play D" strategy.
I think end of game situations would really need to be accounted for in the UFA data. End of quarter possessions seem like lost conversions for the offense and could be very impactful. Additionally winning teams are likely wasting time on offense at the end of game rather than scoring. This lead was generated by breaks so conversion on D could be negatively correlated with conversion on O. One last point, theres often timeouts on defense with a whole sale sub for offense. That would also be interesting to parse out. Really appreciate the look at the data, even though it doesnt say much on the surface, I think theres a lot of interesting info there.
I did mention the EOQ buzzer beater situations in the post ("It’s hard to compare UFA and Men’s club directly since I’m not sure if the UFA data excludes buzzer beater attempts.")
The EOQ time wasting situations I admit I didn't think of, so, good note! I wonder if I can get some info from the UFA on how those are included in the posted conversion rates.
I'm also not sure how it's handled when the O-line is subbed on after a timeout on a D-point. I feel like I heard someone say once that that actually IS recorded as an O-line point. But I'd have to actually find a source before I could be certain that's right.
But it's worth noting both of these situations happen to D-lines as well (though perhaps not as much, certainly). But sometimes D-lines also kill clock after getting a turnover without much time left. (Or chuck up a buzzer beater when the opposing offense turns it with a few seconds left)
For sure those would mitigate the impact to O lines cause we’re looking at differences. Appreciated that you mentioned the EOQ cause it inspired some of my thoughts. Some of this reminds me about how football data sets need to be curated to get rid of non relevant situations. You’re doing cool work looking at the available ultimate data (thanks for pointing these sources out). Keep sharing it’s making the spectating more fun.
Cool to see the conditions of Rockford show up so starkly in the data.
I think getting O & D-line conversion rates for club nationals would be huge in having a better understanding of how teams performed, answering questions like:
- How likely was the o-line to score each point?
- How likely was the o-line to turn it over, or get it back?
- How likely was the d-line to punch it in?
- How likely was the d-line to generate a turnover each point?
- Which teams were the best at generating blocks on top tier o-lines?
- Which team's o-lines were the best at holding against the best defenses?
I believe there is sufficient data feeding into the USAU Stats App to answer these questions but it's not easy to get out (for each nationals game it shows point by point assists, goals, turns, with player names/teams tagged, so you can infer if it's an o or d point).
Anyhow, I think Stacey Waldrup would be the person at USAU to reach out to if you ever wanted to dig into this. Happy to connect y'all if that might be helpful / of interest.
I went and found the app so you don't actually have to answer that question! Yeah I had literally never seen that app before. And yeah that could definitely be used to get more accurate Oline/Dline conversion rates.
I generally prefer using websites to using apps, so if that's what you're saying it's legitimately news to me. And yes it could be very useful for expanding on the data in this post.
Also if you check out the scoring progression table at the bottom of the website page, that could also be used to determine whether it's an O-point or D-point. You'd just need to know who pulled first to determine the O/D status of the first point of each half.
some more info/context for rhino data:
You have the O and D possession-based conversion percentages above. On a per-point basis, our O line had an 84% conversion, while D line had 78% (on points where they got a turnover). I think what this means is that the O line was better at getting the disc back after a turn, while the D line, if they got a chance and turned it, generally didn't get another one. This is borne out by another stat, which is that on points where the D line got a turn, they averaged 1.11 possessions. So it was pretty rare that they got more than one possession. That all makes sense to me since a multi-turn point becomes very tiring/taxing on the defense.
Obviously we went on a great run that tournament, and our O line's ability to get the disc back and D-lines ability to punch in breaks was a big part of it. Our efficiency in the bracket was also much higher overall.
I think a significant factor could be field position at the start of the possession. O-line possessions are likely to start from the pull around the brick or front of the end zone they are defending. Whereas defensive possessions are more likely to be further down the field after getting a turnover. Since the defence tends to have less distance to go before scoring, they are more likely to score before turning over the disk.
Yeah that's a good point. IDK if there's any USAU data that has field position, but UFA data could definitely be used to investigate that hypothesis.
I think the other big factor in my mind was that D-lines are playing offense against units that mostly practice playing offense (O-lines) while O-lines are playing offense against units that mostly practice playing D (D-lines).
So D-lines might not be stronger playing offense "in a vacuum", but they do sometimes have the relative advantage due to the frisbee meta of putting O-focused players on O-lines
This makes me wonder, how important is the starting position for when the defence starts their offensive possession? Would teams that get more D's on under & handler space throws have a higher conversion percentage than teams that get most of their D's on deep shots? This is similar to asking how effective is the "punt and play D" strategy.
I think end of game situations would really need to be accounted for in the UFA data. End of quarter possessions seem like lost conversions for the offense and could be very impactful. Additionally winning teams are likely wasting time on offense at the end of game rather than scoring. This lead was generated by breaks so conversion on D could be negatively correlated with conversion on O. One last point, theres often timeouts on defense with a whole sale sub for offense. That would also be interesting to parse out. Really appreciate the look at the data, even though it doesnt say much on the surface, I think theres a lot of interesting info there.
I did mention the EOQ buzzer beater situations in the post ("It’s hard to compare UFA and Men’s club directly since I’m not sure if the UFA data excludes buzzer beater attempts.")
The EOQ time wasting situations I admit I didn't think of, so, good note! I wonder if I can get some info from the UFA on how those are included in the posted conversion rates.
I'm also not sure how it's handled when the O-line is subbed on after a timeout on a D-point. I feel like I heard someone say once that that actually IS recorded as an O-line point. But I'd have to actually find a source before I could be certain that's right.
But it's worth noting both of these situations happen to D-lines as well (though perhaps not as much, certainly). But sometimes D-lines also kill clock after getting a turnover without much time left. (Or chuck up a buzzer beater when the opposing offense turns it with a few seconds left)
For sure those would mitigate the impact to O lines cause we’re looking at differences. Appreciated that you mentioned the EOQ cause it inspired some of my thoughts. Some of this reminds me about how football data sets need to be curated to get rid of non relevant situations. You’re doing cool work looking at the available ultimate data (thanks for pointing these sources out). Keep sharing it’s making the spectating more fun.
Cool to see the conditions of Rockford show up so starkly in the data.
I think getting O & D-line conversion rates for club nationals would be huge in having a better understanding of how teams performed, answering questions like:
- How likely was the o-line to score each point?
- How likely was the o-line to turn it over, or get it back?
- How likely was the d-line to punch it in?
- How likely was the d-line to generate a turnover each point?
- Which teams were the best at generating blocks on top tier o-lines?
- Which team's o-lines were the best at holding against the best defenses?
I believe there is sufficient data feeding into the USAU Stats App to answer these questions but it's not easy to get out (for each nationals game it shows point by point assists, goals, turns, with player names/teams tagged, so you can infer if it's an o or d point).
Anyhow, I think Stacey Waldrup would be the person at USAU to reach out to if you ever wanted to dig into this. Happy to connect y'all if that might be helpful / of interest.
I went and found the app so you don't actually have to answer that question! Yeah I had literally never seen that app before. And yeah that could definitely be used to get more accurate Oline/Dline conversion rates.
Wait are you telling me that there's a USAU app that has different info than what's posted on the website? (eg https://play.usaultimate.org/teams/events/match_report/?EventGameId=mSQqHP3tW2ag2QBFsid3dTbOj4ZFv4y7d7r3x%2fcrnAo%3d )
I generally prefer using websites to using apps, so if that's what you're saying it's legitimately news to me. And yes it could be very useful for expanding on the data in this post.
Also if you check out the scoring progression table at the bottom of the website page, that could also be used to determine whether it's an O-point or D-point. You'd just need to know who pulled first to determine the O/D status of the first point of each half.
Yeah there’s a bit more in the app - not sure how to post a screenshot here. Can email you if you lmk your email (I’m at Nick.spiva@gmail.com)