Bash Mo Here

DoD and Mango,

you have got yourselves a deal.
But if he doesn't average 6.5 rebounds for the year the opposite holds.....

Dod you will have to do the head up the arse thing and Mango who can continue to be the only one not to believe my CB and cave story..

Smeg

------------------
"It's incredible to see how he [Mobley] can beat his guy one-on-one every
time. That is a tremendous talent. The combination of Steve and Cuttino, I
think, is wonderful in the backcourt."
--the Dream

 

thanks partner!

------------------
It just goes to show how skewed our priorities are when Mo Taylor makes millions of dollars while some high school teacher, that can actually rebound, scrapes to make a living.

If Mo were half the power forward that Charles Barkley was, he'd be 3'2" and still board more than he does now.

 

Just a place to start...

TABLE iso:
[*]player TEXT(20)
[*]game NUMBER
[*]isolation NUMBER -- (sequential number)
[*]success NUMBER -- (0 or 1)
[*]comments TEXT(255)

SELECT player, COUNT(*) PLAYS, AVG(success) SUCCESS_RATE
FROM iso
WHERE game = 1
GROUP BY player

-- leave out the where clause for season numbers


------------------
"We end up with Francis/Mobley isolations very often. More often, I'd say, then other teams." -- jamcracker

[This message has been edited by Tb-Cain (edited November 21, 2000).]

 

good start tb-cain but the success, given rudy's supposed "%ages" will have to be blown up into a table itself w/ different enumerated values.

------------------
It just goes to show how skewed our priorities are when Mo Taylor makes millions of dollars while some high school teacher, that can actually rebound, scrapes to make a living.

If Mo were half the power forward that Charles Barkley was, he'd be 3'2" and still board more than he does now.

 

That's between Rudy and his God. I'll stick with yes or no, myself.

Another thought is to add a field HOME NUMBER (0 or 1) so reports could be generated for success at home versus on the road. In my mind, that's overkill, but to each his own.

------------------
"We end up with Francis/Mobley isolations very often. More often, I'd say, then other teams." -- jamcracker

 

We ought to move this SQL database talk out of the Game Action forum.

To where? Maybe cc.net feedback? Or hangout?

I wanna make a more generalized system that would allow me to calculate success rates for more plays than just ISOs. I just started calculating the ISO rate first because its the easiet play to see (and we do run it more often than most other teams, I'd say). I'd also eventually like to let other people rate plays in the same database and be able to both separate and aggregate different users' play success ratings.

Here at the office we always try to make highly normalized SQL tables. Maybe you'll see what I mean by normalized by looking at my tables.

My first pass at SQL tables:

table USER (the different users who are rating plays)
[*] userID int
[*] username char(20)

table PLAY (the different plays the Rockets run and users rate)
[*] playID int
[*] name

(examples of play names: "Mobley ISO", "Francis ISO", "Mo ISO", "Wizard double pick", etc.)

table POSSESSION (some Rocket possession in a given game)
[*]instanceID int
[*]gameID int
[*]quarter int
[*]secondsLeft int

table OUTCOME
[*] outcomeID int
[*]result: made shot/missed shot/turnover
[*]NULL shooter char(20) (the player who took the shot; might be NULL)
[*]NULL turnoverer char(20) (the player who turned over; might be NULL)
[*]NULL assister char(20) (the player who got the assist; might be NULL)
[*]NULL story text (a text story saying what happened; might be NULL)


table USER_PLAY_POSSESSION_OUTCOME_MAP
[*]userID
[*]playID
[*]possessionID
[*]outcomeID

Now, I'm watching the game and I see a Rocket possession and figure out that we ran an ISO for Mobley. I create a POSESSION row to capture the game and time I saw the play. I choose the play (Mobley ISO) from the options in the PLAY table. I create a new OUTCOME row and enter what happened. Then I create a USER_PLAY_POSSESSION_OUTCOME_MAP row that will map my USER row to the PLAY row I chose, to the POSSESSION row I created, to the OUTCOME row I created.

Of course, now everyone says "Why on earth do you want to use 5 tables? Why not combine the OUTCOME, POSSESSION, and PLAY tables". Or maybe they say "Why don't just put everything in one table?" I'm not going to argue about normalizing SQL tables; I'll just say that normalizing tables has saved me and my fellow developers DAYS of work when we had to change the structure of SQL tables after they were full of data.

Oh, I just saw Tb-Cain's new post, so I'll add this table

table GAMES
[*]gameID int
[*]opponent char(20)
[*]location char(20)


[This message has been edited by jamcracker (edited November 21, 2000).]

 

IT HAS RETURNED!!!!!!!!!!

Let's get this thread to 1000 posts!!!!!

Oh yeah, stick to the subject... Mo sucks.

------------------
Jazzkiller