Fun V. The Weight Of Expectations

[veivoli]

I'll participate. It'll be a fun thing to do. However, after thinking about it for about 30 seconds, it could get a little messy. Not all substitutions would be a "new" combination. Some substitutions will be a previous used combination. The easiest example: the original starting 5 coming out to start the second half. Counting just the unique combinations, and counting them just once, won't make the job any easier.

Is this meandering off track just a wee bit?
I've been in IT for about 50 years, starting in the days when mainframes roamed the earth, in a wide variety of roles. Great career for someone who studied physics and maths...
Living out here, a very few thousand kilometres from the ice wall surrounding our flat earth[1], we don't get much live coverage of our team's games. So I'm often listening to the audio stream from the Varsity Network while watching live stats from StatBroadcast.
Putting those together, it should be possible to develop a database[2] system to record the substitution log, so we'd know the on-court 5 at any time, to the minute. Analysis of the database would allow us to calculate the different combinations, how many time each had been used, and how long each combo had been on court.
The log only records substitution time to the minutes...
However, it might have to be real-time data acquisition. I've had a look at the archive of the BC game on the 14th (my time - 13th for the USAians) and only the last half dozen or so subs are recorded. This adds some complication as a screen scraper would have to be developed to capture subs as they happen.
Maybe someone has done something similar, as I'm sure I've seen charts showing who is one the floor in posts here. If someone can point me to a source for these I'll see if there's a way to do the combination analysis...

Feetnote:
[1] No correspondence shall be entered in to. Put this together with the subjects I studied in depth and work out for yourself if this is a joke - or not.
[2] This is the type of task that people often try to perform with a spreadsheet. A good way to tie yourself in knots!

 

[BBallF]

Isn’t there a stat service that already tracks this information? Her Hoop Stats or someone like that.

I know it tracks it for individual players but I don't know of one that does different combinations on the floor as a whole. If someone does, that would be very useful information.

 

[BBallF]

I will simply make list of the different "fivesomes" on the court during the Louisville game. We are only counting different combinations at this time. It may take a few games to get to 3003! We can get into analytics a little deeper when we decide who is playing for us in the Final Four.😂

 

[FanSince17]

Is this meandering off track just a wee bit?
I've been in IT for about 50 years, starting in the days when mainframes roamed the earth, in a wide variety of roles. Great career for someone who studied physics and maths...
Living out here, a very few thousand kilometres from the ice wall surrounding our flat earth[1], we don't get much live coverage of our team's games. So I'm often listening to the audio stream from the Varsity Network while watching live stats from StatBroadcast.
Putting those together, it should be possible to develop a database[2] system to record the substitution log, so we'd know the on-court 5 at any time, to the minute. Analysis of the database would allow us to calculate the different combinations, how many time each had been used, and how long each combo had been on court.
The log only records substitution time to the minutes...
However, it might have to be real-time data acquisition. I've had a look at the archive of the BC game on the 14th (my time - 13th for the USAians) and only the last half dozen or so subs are recorded. This adds some complication as a screen scraper would have to be developed to capture subs as they happen.
Maybe someone has done something similar, as I'm sure I've seen charts showing who is one the floor in posts here. If someone can point me to a source for these I'll see if there's a way to do the combination analysis...

Feetnote:
[1] No correspondence shall be entered in to. Put this together with the subjects I studied in depth and work out for yourself if this is a joke - or not.
[2] This is the type of task that people often try to perform with a spreadsheet. A good way to tie yourself in knots!

Like baseball analytics now. They can report in real time, a batter's stats in a hundred different ways; vs type of pitch, for each count, against each pitcher, in what inning.......
Sounds like a great adventure for an IT type. (or a 5th grader- I'm a teacher)
Yankee fans- did you catch any games Joe Girardi broadcasted. I like him but analytics not so much. I think a keen eye can catch a lot of stuff. Lotta data rattling around in our heads + other intangible information.

 

[veivoli]

I know it tracks it for individual players but I don't know of one that does different combinations on the floor as a whole. If someone does, that would be very useful information.

I used to watch the Irish guy until I couldn't stand his bias any longer. I'm sure he used to put up charts that showed a team's scoring with different combinations, but that isn't quite what I'm thinking of...

 

[Dogdude]

Now this is not thread drift, it's genetic mutation!

Sssshhh...it's keeping people busy. No telling what thread might crop up if this one stays on course.

 

[veivoli]

However, it might have to be real-time data acquisition.

I are idiot. It's simply a matter of downloading the play-by-play for each quarter and extracting the subs from there. The starting five is in the game box....

 

[Skeets]

Now this is not thread drift, it's genetic mutation!

I'm sorry. This thread was supposed to be discussing your neurosis.

 

[RockyMTblue2]

I'm sorry. This thread was supposed to be discussing your neurosis.

Now, that's a great reply. I'll be chuckling all day.

 

[Bone Dog]

I will volunteer to keep track for the Louisville game and make it a thread after a sufficient number of replies (probably in excess of 100) to oldude's post game thread have taken place. It would be even more interesting to see how the 3003 different combinations fare in terms of + - points in a given game. I am sure one of our talented algorithmic BYers can figure that out. I think the pool should be how many combinations are predicted for an upcoming game. For example, 10 in the Louisville game? 15? 25? Remember each substitution is a new combination.

I don't know this will help, but this is one of the services I was trying to remember earlier: Hoop Explorer. It offers cumulative and per game lineup analyses.

https://hoop-explorer.com/LineupAnalyzer?baseQuery=&filter=&gender=Women&maxRank=400&maxTableSize=50&minPoss=5&minRank=0&sortBy=desc%3Aoff_poss&team=UConn&year=2024%2F25&