What up Nerds! Back at it again with another 'Advanced' Box Score, but doing it a little different this week. Some of the feedback in previous threads was that there was less interest in trying to use metrics to better understand what happened in the game, and more interest in understanding how a given performance compares to past performances. So, to do this, we have 6 'advanced' comparative box scores.
When VT is on Offense:
First, let's compare how this VT offensive performance compares against the season's past offensive performances
The five right-most columns are the Min, Max, and different quartiles for all 9 games. The color-coded columns are VT performance against Syracuse. Green is good and red is bad (relative to past performances). As you can see, VT had season highs in Success Rates and most rushing statistics.
Next, we can compare how the VT offense performed relative to Syracuse's past opponents
What we see here is that no other Syracuse opponent was able to run on Syracuse quite like VT did.
When VT is on Defense:
Here, we are comparing this VT defense performance against the season's past defensive performances
Lots of green means that this was the best defense performance (NOT adjusted for quality of opponent) this year.
Now, we can compare how the VT defense performed relative to Syracuse's past opponents
Here, we see that this was one of the worst offensive performances that Syracuse has had to date
Comparing this VT performance to National Averages
Okay, so this comparison is a little wonky because we're comparing VT's performance in a single game to the average performance for each other team.
For example - take a look at Total Line Yards per Game. You'll notice that VT got 203 Line Yards against Syracuse, but 'Maximum' for the country is 200. That's not saying that VT has the most like yards in a single game all season; rather, it's saying that Arizona is averaging 200 line yards per game, and that is the most any team is averaging this season.
As always, data comes from Collegefootballdata.com. If any questions, check out their glossary.
Let me know if you like this format. Questions and Feedback are welcomed.
Comments
A whole lot to digest there.
Good feedback - is it just an overwhelming amount of numbers or is it presented in a way that isn't clear? Do you prefer the previous format?
I prefer this format. I find it is easier to read, though it did take me a bit to figure it out. A sightly more elementary and long (but recyclable) explanation might help.
maybe I was just used to it but I liked the original format better - there was more information and I could kind of piece things together myself. This format feels like less info but I haven't spent much time with it - I probably just need to get used to it again
It was inspired by this comment from other Ben:
To oversimplifying the OP:
But, if this is overwhelming, maybe I keep the format that I used in the past, and just add a second table that details our 'typical' performance in each area?
I liked Ben's comment. I think there is value in comparing how we've done against a specific opponent versus how we've done generally in the season as well as how we've done vs how their other opponents have done. I think I just need a couple weeks to get used to the format
The color code in the left column helps a lot.
For a guy like me that doesn't look at info similar to this very often, i have to consider what it means perhaps a bit more than most.
I see graphs more often.
Your data is outstanding.
happy the color coding helps.
Let me know if you feel anything else is overwhelming/confusing/etc. happy to adjust stuff in the future.
I have maybe absorbed 27% of this but great work!
I know this is about cuse but I saw FSU's October defensive numbers and they are brutal, not 1 passing TD allowed, less than 41.7% completion (Drones was 10 points above), 4.2 YPComp (drones was well above that). So we awesome compared to other teams against FSU.
Really interesting! Our success rate was absolutely stellar on both sides of the ball. Passing game still has some work to do with the exception of hitting a couple of explosive plays per fame. And this seems to be the archetype of defense we are used to. No bending, with the occasional break/glitch (great success rate, but some explosiveness conceded).
Our success rate against Syracuse was amazing.
Re passing: I'm trying to figure out if there are any unique stats to passing (similar to line yards, second level yards, etc). Some things I've read about that I want to learn more about is average depth of target (ADOT), YAC, etc.
Looks like we kicked their ass harder than anyone else. That agrees with what I saw watching the game. It shows a trend of improvement. I hope we keep that up.....
Bar - I'm wondering if it's possible to look at a moving average of, say, our last 4 games, potentially weighted towards the most recent. Maybe also weigh FSU higher as a better opponent? Just thinking that our average is skewed bad by our early games. Also not sure how much better our early opponents were, hence weighing FSU higher, while trying to also look at improvement in our performance over time when comparing to national average.
Just some thoughts - I'd be curious to see what those comparisons would show.
Edit: comparing one game against past performance seems like it could be skewed, so using more than one game might be helpful, and maybe instead of 4 games like comparison of last 2 games against first 2, or last 3 against first 3. So, like I said, using recent average to remove highs against a particularly weak opponent, but also looking for a more marked view of improvement over time considering we looked like we turned a corner where performance after that point in time might be markedly different than that from before. Hopefully that makes sense.
It is possible... it's such a small sample size, I'm not sure how useful it would be, but I'll take a look tomorrow or Friday.
That is not possible in this analysis. Nothing is weighted or opponent adjusted.
My apologies - I didn't go look at the source to see what the site provided vs calculations you performed.
I just like looking at data in different ways and especially looking for trends and how to explain the trends.
I do usually like numbers, but to what I'm finding is a large extent, I just want to be able to watch and enjoy the games in the moment. Along those lines, I have found myself thinking that even though we beat ODU, we just LOOK so much better in how we perform.
I've also been pretty worn out recently, so I have insufficient brain energy required for detailed analysis. Or so my brain says...
I also don't know that a moving average is going to be very informative as the data is highly matchup dependent and a moving average is typically used to track performance across a timeline of similar conditions
Circling back to this... I filtered out all of the OOC games, and the results look basically the same:
Our performance against Syracuse was our most efficient all season. Compare the single game offensive stats to their respective maximum, and the defensive stats to their respective minimum and you'll see a lot of overlap.
Another thought - I'm wondering how much of our "poor" performance on offensive passing plays was just because we relied a lot more on rushing this game (because we were having so much success). I mean, we had 528 yards of offense, 318 of which were from rushing).
And our early season performance was probably skewed towards passing with Wells as the QB and limited effectiveness of rushing early. (If I'm thinking about this correctly).
Curious - what leads you to believe our passing performance was poor?
Drones only hitting on 2 of what 7 long bombs?
Distilling the passing offense to just long balls seems a bit reductive, no? 61% pass completion overall, but if you ignore the 2/7 (I'm just assuming this stat is correct, not hyperbole) on deep ball, drones had a 73% completion percentage (14/19).
Big different between 'this team is bad at passing' and 'this team is bad at the deep ball'
AND deep shots are generally going to hit less. There's a higher degree of difficulty on those throws. Elite QBs complete 40-50% of their deep shots.
I would contend that a bigger indication of the passing game's limitations is how infrequently we target the middle of the field with standard drop back passes. I'd be very curious to see Drones's passing chart because it's likely decent near the line of scrimmage and deep, but not much over the middle.
It felt like a couple of those deep shots early were just there to make sure Cuse didn't pack the box. It would be great to hit them, but just showing them we are going to take a shot keeps them honest.
Yep - I read/heard somewhere that Syracuse's defense is very aggressive, so to my ignorant ass, it makes sense that Bowen would dial up a couple deep balls (to keep the secondary focused on receivers) and the trick plays (which would freeze the middle layer of the defense), both of which would give our ground game more room to breath.
Nothing will slow down an aggressive corner trying to tackle an RB to the outside like said RB throwing a TD over his head...
The two red items for passing on our offense. Not that it was "poor" exactly just that we had the two red items (probably should have clarified rather than just putting it in quotes (which, for me, means I'm using the term loosely)). Without going back and looking at passing attempts from previous games, just wondering if we were passing less because of how effective we were rushing, so it may have skewed the stats. Not that we were "poor" since we had 2 passing TD's and over 200 yards passing as well - just not as "good" compared to previous games.
Maybe also a realization that a red item isn't necessarily bad. To use a little hyperbole to illustrate the point. Say we pass once and it's an incompletion, but we put up 600 yards rushing with 5 rushing TDs, passing would probably look really bad compared to previous games, but it didn't necessarily mean anything bad in that particular game because we were able to rush all over our hypothetical opponent.
Hopefully that makes sense - just trying to rationalize why we were seeing the red items for passing.
Looks good, color coding definitely helps with the analysis and quick uptake. I think you could simplify with only the 1Q, mean, and 3Q. The max and min are likely atypical numbers from matchup strengths and weaknesses lining up, garbage time, etc. and add less value to the total data set. Eliminating them would chop down the information load by 40% without sacrificing much performance assessment capability.
This is great feedback, Thank you!