QB completion percentage

[Knighty Knight]

Can someone explain the maths to me? I would like to understand why Bortles had a downward adjustment of 15% compared to his real life completion percentage.

Blake Bortles (41%:40)
Russell Wilson (60%:60)

CRUNCHING THE NUMBERS: ADJUSTMENTS TO WEEKLY UPDATE STATS
Tennessee Offensive Line Grade (5.6) vs. Miami Front Seven (6.4) = -0.8 in the trenches
Miami Offensive Line (5.9) vs. Tennessee Front Seven (6.4) = -0.5 adjustments to running averages

Tennessee QB Completion % Impacts (x.1): WR (7) + OL (5.6) - Miami F7 (6.4) - SEC (6.5) = -0.03
Miami QB Completion % Impacts (x.1): WR (6) + OL (5.9) - Tennessee F7 (6.4) - SEC (7.3) = -0.15

Tennessee QB1: update 19-for-33 (58%, -0.03 adj) 298-yds (15.7 ypc) 3/2 | SIM: 24-for-41 (59%) 350-yds (14.6 ypc), 2 TD / 2 INT
Miami QB1 update :18-for-33 (55%, -0.15 adj) 176-yds (9.8 ypc) 1/1 | SIM: 16-for-41 (39%) 173-yds (10.8 ypc), 1 TD / 2 INT

[Goodell]

Can someone explain the maths to me? I would like to understand why Bortles had a downward adjustment of 15% compared to his real life completion percentage.

I may not be fully understanding the detail in question, but generally speaking that's how our sim works by design. We don't just give players the exact same stats they have in real life, but they are ADJUSTED based upon the supporting cast compared to the opposing defense.

If a QB was 55% in real life, but his team had a -15% disadvantage
[Miami QB Completion % Impacts (x.1): WR (6) + OL (5.9) - Tennessee F7 (6.4) - SEC (7.3) = -0.15]

Then the sim would be working toward a 55-15 = 40% completion percentage. Actually it's an even bigger disadvantage in the grades (-18) but the system caps max disadvantage to 15.

[Knighty Knight]

That's what I gathered when looking at Bortles but wasn't sure due to the inverse correlation with Russell Wilson.

Bortles completion % resulted in a 1:1 correlation between real life % - 15% disadvantage. I expected Wilson see a 3% decrease from real life, not a 2% increase for a net 5% increase based on -03% disadvantage.

[Goodell]

That's what I gathered when looking at Bortles but wasn't sure due to the inverse correlation with Russell Wilson.

Bortles completion % resulted in a 1:1 correlation between real life % - 15% disadvantage. I expected Wilson see a 3% decrease from real life, not a 2% increase for a net 5% increase based on -03% disadvantage.

Yeah, and it won't always work out exactly in the end because of how the sim works. A QB doesn't stay at their same percentage all game long. Basically if it's working toward 40% for example and a QB is at 35% then the odds go way up for a next pass completion, but odds going way up isn't guaranteed completion. If a QB is way above their expected percentage, then the odds for next pass incompletion go up. There's also auto-TDs in the red zone, so for example even if a QB was above their percentage but the team gets in the redzone and they have a TD in their update still, then it's a TD pass regardless of completion percentage status. But then that QB would be way above percentage so next series the odds work toward more incompletions to work toward the goal. Sometimes the game stops or no more throws before it gets a chance to work toward correction.

In that Wilson case, it's setup to work toward a 55% for him (58% in reality with -3% adjustment).

You'll see early in the game it's doing just that working toward a 55%. It's recorded in the play-by-play: Russell Wilson (4-for-7 58 yds | 57%:55)

In the second half, you'll notice it shifts on Wilson to work toward a 58%. I'll have to go back in the programming but I'm pretty sure that's because he's behind on his projected TDs so in the second half of games if behind TDs it boosts the completion percentage slightly to try to help that QB behind his TDs. Russell Wilson (13-for-23 199 yds | 57%:58) And I believe it might boost a little again in the 4th if still behind TDs. Early on the the sims development years ago people complained so much about QBs not getting their TDs so we boosted them a little as the game goes on if behind TDs.

Toward the end of the game, if a QB goes beyond their game update and is getting more sim throws than their game update, then it reverts to a default percentage for a QB of their grade the rest of the way for throws beyond the game update. So if you had a QB with only 20 throws in their NFL game update but you sim situations called for lots of throwing in a shootout and your same sim QB had 50 throws, the last 30 would be based upon default update instead of the NFL game completion percentage for a small sample. The best reason for that is cases where an NFL QB might throw a small amount like going 5-for-5 filling in a game in reality but we don't have them throw 100% for 50 throws in the sim. They get that 100% for the first 5 throws of their update then it goes to default percentage for their grade for throws beyond that.

Hope that helps. Appreciate the chance to describe that, especially to newer GMs.

[Knighty Knight]

Makes sense. Just wanted to ensure I understood the sim logic. Thanks for the clarification!

[splitsinker]

Along the same lines, I dont understand Cam Newton's update this week at all.
Instead of going up 5%, he went down 6%

Oakland QB Completion % Impacts (x.1): WR (7.3) + OL (5.8) - St. Louis F7 (6.2) - SEC (6.4) = 0.05

Oakland QB1 update :17-for-26 (65%, 0.05 adj) 161-yds (9.5 ypc) 0/0 | SIM: 32-for-54 (59%) 306-yds (9.6 ypc), 0 TD / 0 INT

thnks for all your hard work commish

[Goodell]

Along the same lines, I dont understand Cam Newton's update this week at all.
Instead of going up 5%, he went down 6%

Oakland QB Completion % Impacts (x.1): WR (7.3) + OL (5.8) - St. Louis F7 (6.2) - SEC (6.4) = 0.05

Oakland QB1 update :17-for-26 (65%, 0.05 adj) 161-yds (9.5 ypc) 0/0 | SIM: 32-for-54 (59%) 306-yds (9.6 ypc), 0 TD / 0 INT

thnks for all your hard work commish

Without looking closer on the surface that looks like the case of throws over the game update I mentioned above. The sim works toward the adjusted percentage up to the number of throws in the update. If he only threw 26 in reality, he'd get that adjusted average for those. All the throws beyond that throwing more than twice as much would be at a default percentage based upon his grade. If we had a QB throw 9-for-10 in the NFL, but the sim has them throw 50 passes instead, we don't have them throw 90% for 50 passes. Just the first 10 up to the game update, then it's at a default percentage with higher graded players getting higher default percentages.

[splitsinker]

that make sense now that you broke it down for me.
thnks commish!
what a cool system!
that 11% swing didn't help our cause, lol. better luck next wk!