2013 Sim results

RebelFan
Posts: 469
Joined: Fri Apr 10, 2009 5:01 pm
Location: MS

2013 Sim results

Post by RebelFan »

I just noticed something from my sim this week that I thought I should bring your attention to commish.
My #2 rb had an update that consisted of 1 carry for -7 yards, which was extrapolated into the sim as 6 rushes for negative 28 yards. Granted I got the adjustment to the ballcarrier's average from my oline, but it just doesn't seem right that 1 negative carry should get multiplied like that throughout the game. It really had an effect on the outcome of my game by killing drives & field position.
I think there should be a 5 carry minimum or something to avoid this kind of update being input.
I could see how it would also make a huge difference if it were 1 carry for plus 7 yards, or one carry for 10.
My point is we should have some sort of safeguard to avoid a 1 carry for 7 yard update becoming a 10 carry for 70 yard sim input...
Maybe if the player has less than 3 or 5 carries we could use a season average or something?
GM - Chicago Bears - AFFL
GM - San Francisco 49ers - DFFL

"Talent Hoarder"
Jared A
Posts: 1130
Joined: Tue Jul 31, 2007 3:18 pm

Re: 2013 Sim results

Post by Jared A »

A problem with that is... what if he had 2 carries for 5 yards but had a TD or a fumble?


I would suggest that negative yards become 0 or -1. To avoid it getting out of hand. But, I don't like the idea of season averages.
RebelFan
Posts: 469
Joined: Fri Apr 10, 2009 5:01 pm
Location: MS

Re: 2013 Sim results

Post by RebelFan »

Yeah, I could see how that would work better.
GM - Chicago Bears - AFFL
GM - San Francisco 49ers - DFFL

"Talent Hoarder"
tino38
Posts: 1132
Joined: Sun Apr 12, 2009 5:39 pm

Re: 2013 Sim results

Post by tino38 »

Is there a glitch in the system? SF in DFFL has major upgrades over Indy. But loses because Kap plays poorly. Miami has a decent team in DFFL and has won 2 in a row without his QB taking a real NFL snap. It feels like the system penalizes a team if their QB struggles, but doesn't do the same if the QB has no negative stats to input for their team...Just curious on this.
BRFL Saints (31-20) (3-0)
- NFCS Champ: 23’
- NFC Champ: 23’
- SB Champ 23’
AFFL Patriots (97-82) (8-4)
-AFCE Champ: 16', 22’, 23’
-AFC Champ: 22’
-SB Champ: 22’
DFFL Jets - SB Champ 21’ & 22’
FFFL Jets - SB Champ 17’ & 18’
Royce R
Posts: 675
Joined: Thu Aug 02, 2007 12:03 am
Location: South Dakota

Re: 2013 Sim results

Post by Royce R »

there is a update giving by grade if you qb didn't play. Then your grades help or reduce that update from there.
AFFL - Titans GM since 2007
96 - 62 - 2 regular season
6 playoff appearances
4 division titles
2 conference titles
1 AFFL title
Goodell
Posts: 3780
Joined: Mon Sep 15, 2008 3:44 am

Re: 2013 Sim results

Post by Goodell »

tino38 wrote:Is there a glitch in the system? SF in DFFL has major upgrades over Indy. But loses because Kap plays poorly. Miami has a decent team in DFFL and has won 2 in a row without his QB taking a real NFL snap. It feels like the system penalizes a team if their QB struggles, but doesn't do the same if the QB has no negative stats to input for their team...Just curious on this.
That's reality isn't it? If a team's QB stinks, they probably aren't going to have great odds of winning. NFL is arguably very QB driven. Great QB play equals improved chance to win. QB stinks and probably in for a rough day.

To me, it kind of mirrors reality. The 49ers have a great team. Their players are high quality. Their talent doesn't change. When QB playing great, they're rolling. When QB not playing well, not so much. Reality and sim. QB's are important in reality and sim.

As mentioned above, when a QB doesn't play in reality, they still get a default update. It's not usually very good. It's either bad if their grade is bad or mediocre if they're grade is good. And from there, grade advantages move that default update up or down like always. Except for strong graded QBs, most of the default updates are not all that good. They are negative and usually (but not always) worse than what most NFL QBs put up weekly. If a QB stinks in reality, though, he may have a worse update than the default guy gets. That's the breaks usually in reality where most teams lose if their QB plays really badly.

For default update, I played a team that started Tim Tebow this week for example. Tebow obviously doesn't start last week so he gets a default update for his grade because it's not against the law for our sim teams to start him, and we don't have him go 0-for-0 just because he didn't start in NFL. He gets a mediocre/bad default update:

10-for-20 (50%, -0.14 adj) 150-yds (15 ypc) 0/1

Most teams wouldn't take that. Most NFL QBs do a lot more than that. But we don't give him automatically 4 INTs every week either. It's just mediocre and not going to win much on his own.

Take the sim grades into account and his default update goes way down by 14% in his matchup against me and adds an extra INT for near max disadvantage. So he throws 36% and has 2 picks. Very bad. However, on the other hand had his team instead had really great grades, and maybe he starts in sim just because of QB injury on a championship level team otherwise, maybe he goes from 50% to 65% with an extra TD in his update if he has a near maximum upgrade for a strong offense.

So the default updates are meant to be pretty vanilla that can sway depending upon the quality of the supporting cast. They are fairly bad on their own but can go to either really bad to fairly good depending upon the grades around them. That's the intent anyway.
Official Statement from the Commissioner's Office
tino38
Posts: 1132
Joined: Sun Apr 12, 2009 5:39 pm

Re: 2013 Sim results

Post by tino38 »

I mean I figured the bad game from Kaepernick would hurt but just figured his grades would over come it. I see where you are coming from as far as a non playing QB getting "vanilla" type stats now. Wasn't sure how that worked. Thanks for response
BRFL Saints (31-20) (3-0)
- NFCS Champ: 23’
- NFC Champ: 23’
- SB Champ 23’
AFFL Patriots (97-82) (8-4)
-AFCE Champ: 16', 22’, 23’
-AFC Champ: 22’
-SB Champ: 22’
DFFL Jets - SB Champ 21’ & 22’
FFFL Jets - SB Champ 17’ & 18’
RebelFan
Posts: 469
Joined: Fri Apr 10, 2009 5:01 pm
Location: MS

Re: 2013 Sim results

Post by RebelFan »

I understand both sides of it and it is very realistic for a struggling QB performance to lead directly to a loss.
I expected to win that game vs. Indy because of the massive advantages in all 4 categories and very good YPC from Gore & Robinson. I also expected Kaep's numbers to improve quite a bit from grade advantages, but it just didn't make much of a difference. His low completion % was (and always is) bad news for any sim team.
The defensive advantages -
Indianapolis Offensive Line Grade (5.6) vs. San Francisco Front Seven (7.8) = -2.2 in the trenches
Indianapolis QB Completion % Impacts (x.1): WR (5.7) + OL (5.6) - San Francisco F7 (7.8) - SEC (6.4) = -0.15
The defensive advantages I expected to carry us to victory, but Peyton was darn near perfect against the pitiful Raiders in real life.
I basically chalk it up to bad luck with Peyton getting all of his tds carried over to the sim (1 int. added in) and Kaep getting all of his turnovers carried over with only one td added. I definitely don't feel like we got cheated or anything like that, just a worst case scenario kind of thing. I have had worse defenses shut down better offenses for the win & I've had it go the other way. That's just the way the cookie crumbles.
Here's the line on the QBs -
Indianapolis QB1: update 32-for-37 (86%, -0.15 adj) 374-yds (11.7 ypc) 3/0 | SIM: 27-for-38 (71%) 331-yds (12.3 ypc), 4 TD / 1 INT
San Francisco QB1 update :13-for-27 (48%, 0.15 adj) 150-yds (11.5 ypc) 0/1 | SIM: 24-for-39 (62%) 219-yds (9.1 ypc), 1 TD / 1 INT
This line is the one that I don't understand.
Indianapolis RB1: 8-62 (6.3 adj) | SIM: 5-70 (14 ave)
That's a rather large spike in average for an RB with maximum oline downgrade, but I see he had a 49 yard run in the sim, which means he probably had one in real life that got incorporated in & threw the average off a bit.

Another minor note still seeing situational things like the Colts throwing the ball twice to run out the clock at the end of the game when they probably should be running.
Also just browsing over it, I can't see where SF used any timeouts in the second half even as Peyton runs the clock out to end the game.
1:31 | 1 and 10 @ 50 (50)... Peyton Manning (26-for-37 329 yds | 70%:71) passes complete to Chase Coffman for 4 yards.
0:49 | 2 and 6 @ 46 (46)... Peyton Manning (27-for-38 331 yds | 71%:71) passes complete to Lamar Miller for 2 yards.
0:00 | 3 and 4 @ 44 (44)...
End of Quarter 4 (Indianapolis 28 to San Francisco 23)
I understand that the updates & grades go into the sim and the sim decides on a winner. So I guess these are cosmetic issues more than anything.
GM - Chicago Bears - AFFL
GM - San Francisco 49ers - DFFL

"Talent Hoarder"
Goodell
Posts: 3780
Joined: Mon Sep 15, 2008 3:44 am

Re: 2013 Sim results

Post by Goodell »

Thanks for the thoughtful comments and questions here. Good discussions.

The biggest reason for surprise upsets is that one roll of the dice. If we'd run a series of a particular matchup and one team would only win 1 out of 10 times, that one time might be the one we sim. I tried, and begged and pleaded the case for a couple of off-seasons to try to reduce upsets and have the game simulator generate more predictable results (running a mini-series itself and then making sure the game it returned to count was the team that won most often), but every time I bring it up it gets voted down. People want, at least in theory, the possibility of upsets like reality. So they will happen at times. I think people mostly like them generally, but hate it when they are the ones who get bit by a random upset popping up. But it's one way or the other. Either much more predictable with less upsets, or sometimes favorites are going to lose and sometimes just because upsets are possible and part of the game.

In that example, I think it's a couple of things. The real-life SF team is very similar. Expected to be elite with some of the best talent, but got off to a slow start in reality because of QB play recently arguably. So that's reflected here also. In one case it was a lot of turnovers I believe. In another case, it was running up against a great QB update on the other side as well. Like the real SF team we saw tonight, there might have been temporary setbacks based upon a string of injuries or turnovers or circumstances, but in the big picture things go as they should most of the time if the QB produces as should be expected behind a strong supporting cast.

For the RB average difference noted, it is the case of that RB having a 49 yard big play in the update. Once that happens, it'll throw the average off for a bit. The sim would try to adjust the average on carries afterward (probably returning some rushes for no yards or negative yards to try to return the average to norm) but those rushes didn't happen for circumstance play calling. He only had 8 in the update as well, so wasn't favored to get a ton of attempts unless the game calling tilted a lot more rushing.

I do need to make some tweaks to the simulation as it's forever a work in progress. I'm trying to tackle one small thing each week that I notice to adjust for next time. Last year we made a more significant change that moved our offenses away from the old 3 yards and a cloud of dust conservatism that was more prevalent when we started this years ago and more toward today's more aggressive passing attacks with more focus on the playcalling toward the QB's numbers. So that tilted some of the playcalling more even in some circumstances where it might not otherwise. But as argued by others last year when something similar mentioned, today's offenses don't always just run to run out the clock any more either. Teams still pass more in those situations than they would have years ago. In most of those cases it's probably trying to get the top players their stats still.

But that's one area I want to focus a little more on perfecting the playcalling and situational details as we go along to improve it overall. It shifted a little too far away from conservative in some cases with recent changes, and may need to rein some of that in with certain situations again. But that's perhaps never going to be perfect though and never the real focus of the game here. EA Sports spends millions and has hundreds of people working on their product that is more focused on actual field play and still gets complaints. I have just me in my spare time and no resources. So this game is more GM team building than game day play-calling oriented. It's a big picture GM game won in the front office more than a coaches or player game on the field. We just want results that generally seem realistic and reward good team building for our FanGMs. So it may never be perfect there, but we'll always keep trying to clean up as we can and improve there.
Official Statement from the Commissioner's Office
RebelFan
Posts: 469
Joined: Fri Apr 10, 2009 5:01 pm
Location: MS

Re: 2013 Sim results

Post by RebelFan »

One interesting thing to look at on that 49 yard play by the RB, is whether he should actually get it in the sim with such a dramatic disadvantage in the trenches.
Another thing I think we should take a look at is removing the .15 cap on grade disadvantages so that the full difference in opposing units can be realized in the sim.
Ie - 7.8 F7 vs. 5.6 oline would get .22 advantage rather than .15 max.
Not sure if that has ever been discussed before but I'm for it.
GM - Chicago Bears - AFFL
GM - San Francisco 49ers - DFFL

"Talent Hoarder"
Post Reply