Flagging this comment for consideration in future developments.

Here the data broken down by offensive formation and passing play distance: https://drive.google.com/file/d/1fSoQuDWfs93H_ZTFQKG_0ommesFfMIj_/view?usp=sharing

Dataset
Data has been agg'ed from six different leagues that IoT and I are active in. These are defensive plays that we have run against our opponents.

The Clean% column is the number of pass attempts against that defensive play minus the number of sacks plus the number of pressures divided by the number of pass attempts.

The position columns are the number of sacks and pressures by a player at that position for the defensive play divided by the total number of sacks and pressures against that defensive play.

Other Notes
The position columns will not always add up to 100% because I'm not agg'ing all positions, and sometimes add up to >100% because I'm parsing the play-by-play text to get the player. So when the play-by-play text has multiple pressures listed (it happens), you get more than one player agg'ed.

Here's a sample of that logic:

CASE
WHEN LDE1 = ''
THEN 0
WHEN Text LIKE '%sacked%'||rtrim(ltrim(LDE1, '#1234567890 '), ' - WRRBTEFBCBFSSSLBLGRGRTLTCDEDTSLBWLBMLBCBSSFS')||')%'
THEN 1
WHEN Text LIKE '%Pressure by %'||rtrim(ltrim(LDE1, '#1234567890 '), ' - WRRBTEFBCBFSSSLBLGRGRTLTCDEDTSLBWLBMLBCBSSFS')||'.%'
THEN 1
ELSE 0
END LDE1

UPDATE: Oh, right, one thing to keep in mind is that the data for the 46 Heavy is skewed badly by one season where I put an 89 SP/AC LB at the LDE spot and hammered off 70 sacks against a highly competitive league.

By raymattison21
3/16/2018 9:05 am
raymattison21 wrote:


But that got me thinking about those dump offs . They seem to be regulated by play call distance ? Long pass ...throw long.... ignore the dump off. Where when you Throw short it ignores the wr and dumps it off. But what should be correct ?

weve been doing this for over two years. A lot of gripes in this thread touch on problems in the code today