Post patch LT regression

[Hooe]

Re: Post patch LT regression

Cross-posting this from the patch impressions thread for visibility and so people understand what exactly is happening / how the regression is occurring:

The issue is that left tackles (and right tackles to a lesser extent, but it's mostly left tackles) are allowing a disproportionately high number of sacks relative to other starting offensive linemen.

The total number of sacks allowed is actually pretty solid at a glance, but teams' left tackles are allowing too many sacks. For example, in one given sim, the Cowboys allowed 32 total sacks, but Tyron Smith allowed 12 of them. Washington Post credits Tyron Smith with six sacks allowed in 2014 and he's never allowed double-digit sacks in his career per that source. Another example: Browns LT Joe Thomas - probably the premier left tackle in the league - allowed 13 sacks in the same simulated season (with PBK of 96 following the season); he's never allowed more than six sacks in his career, per Washington Post.

The extreme example: for the Rams, they allowed 60 sacks as a team in one sim I did, and Greg Robinson allowed 41 sacks by himself. No other player on the team allowed more than 9.

Combine this with a component of Game Prep where an offensive tackle will automatically regress if he allows too many sacks in a game. For example, Greg Robinson will automatically lose four points in Pass Blocking permanently if he allows four sacks in a game. The amount of regression may be tied to Consistency; the Rams' right tackle, Rob Havenstein, also allowed four sacks in a single game once this sim, and he lost five points on Pass Blocking instead of four. It appears that three sacks allowed is the threshold for any ratings penalty, so an OL can allow 0, 1, or 2 sacks in a game and be fine.

Going back to Greg Robinson allowing 41 sacks in a season - that's a little over 2.5 sacks a game. In a second sim I did, he allowed at least 3 sacks in 8 games, and in two games allowed more than 6 sacks individually. So he's getting hit with PBK ratings penalties at least eight times in the season.

That's the reason we're seeing the tackles regress so much - they are allowing a ton of sacks individually, and the game is penalizing them for that.

For frame of reference - Washington Post credits Vikings LT Matt Kalil, one of the worst current starting tackles in the NFL by consensus, with 13.75 (.75?) sacks allowed in 2014. Unfortunately sortable real-world individual OL stats are proving really hard for me to find.

If I were to take a stab at what the long-term effects of this would be - CPU teams are going to be at the mercy of the sim engine with respect to developing young offensive tackles, and any who don't perform in the game sims are going to probably discarded quickly on account of the extreme PBK regression. The LTs who already have high ratings, though allowing too many sacks, will probably maintain their ratings more often than not because they mostly won't ever hit the single-game sacks allowed thresholds to incur ratings penalties.

[Jr.]

Re: Post patch LT regression

Great work, CM. Yet another mishap from the backward progression/regression system that's based on stats/XP. Guys don't get worse from giving up sacks... they give up sacks because they're bad. If the game ever gets this kind of mindset (that stats are a reflection of ability, not the other way around) a lot of these issues can be remedied.

[rspencer86]

Re: Post patch LT regression

Great work, CM. Yet another mishap from the backward progression/regression system that's based on stats/XP. Guys don't get worse from giving up sacks... they give up sacks because they're bad. If the game ever gets this kind of mindset (that stats are a reflection of ability, not the other way around) a lot of these issues can be remedied.

YES. I'm glad to see I'm not the only one confused by this line of thinking.

Ratings should drive performance, not the other way around.

[Robo COP]

Re: Post patch LT regression

Good find guys. One question though

Are LTs on teams with QBs that have decent sense pressure doing okay? I know there are a couple of different sense pressures from what I recall (isn't there like ideal, oblivious, and some others?) Just wondering how somebody who fits in the middle of that would fare

[kehlis]

Re: Post patch LT regression

Great work, CM. Yet another mishap from the backward progression/regression system that's based on stats/XP. Guys don't get worse from giving up sacks... they give up sacks because they're bad. If the game ever gets this kind of mindset (that stats are a reflection of ability, not the other way around) a lot of these issues can be remedied.

YES. I'm glad to see I'm not the only one confused by this line of thinking.

Ratings should drive performance, not the other way around.

This XP system just has to go. It really makes absolutely no sense in a sports game of any kind.

[PhillyPhanatic14]

Re: Post patch LT regression

This XP system just has to go. It really makes absolutely no sense in a sports game of any kind.

How did they release a patch that doesn't fix anything? Did they not test it at all before sending it out to us? ...much like what they did with the game.

[BleedGreen710]

Re: Post patch LT regression

dang.thought I could finally begin a new franchise

ah well

really hope this gets fixed soon, thanks for posting your findings

[Hooe]

Re: Post patch LT regression

How did they release a patch that doesn't fix anything? Did they not test it at all before sending it out to us? ...much like what they did with the game.

It did fix a number of things. Unfortunately CFM just has plenty more things to fix.

[Sajiky]

Re: Post patch LT regression

Anyone do any testing to find if there are any possible slider settings to get this problem at least mitigated enough to get a cfm started?

Been waiting long enough to get a career going.

[bad_philanthropy]

Re: Post patch LT regression

It did fix a number of things. Unfortunately CFM just has plenty more things to fix.

This. EA did a good job addressing things quickly, but CFM is a pretty bloated and complicated old heap of code at this point. It took them years to iron out things like younger, lower rated receivers starting over established players, and wonky sim stats back in the PS2 era. These days CFM has so many extra variables heaped on to the core sim engine it's a wonder it can spit out anything resembling the NFL at all.

I think we need to understand the state CFM is in and try to make the best out of it this year. I accept Rex's admission that it needs work, and will wait see what Kolbe and Rex do with the mode for next year. It's just the reality of the process. Gameplay needed so much attention, and I'm glad this regime focused on it the way they did. It's as good as it's been in forever.

I said it in another thread, I think the Madden team should look long and hard at NHL's Be a GM mode from Gen 3 and now in NHL 16 as a model for player development, progression, and regression.

[EricFreakingBerry]

Re: Post patch LT regression

It's frustrating that last year the sim engine couldn't produce any sacks....so to offset that....this year they jack up the sim stats for sacks but unevenly distribute it against the tackles. I rarely see centers ever giving up sacks even though a few dominant DT's like Suh and McCoy are racking up sack numbers. I assume the sim engine lined them up against the tackle for all their sacks.

Based on what I can tell though, I don't think 10-15 year degradation is going to be *that* bad because from what I can tell in testing the drops occur mostly to the top 4-6 tackles that give up the most sacks, and as long as the draft brings in about 4-6 tackles that are decent pass blockers every year then it shouldn't fall apart into a land where there's nothing but 50-60 OVR tackles.

[drustrk3]

Re: Post patch LT regression

It's frustrating that last year the sim engine couldn't produce any sacks....so to offset that....this year they jack up the sim stats for sacks but unevenly distribute it against the tackles. I rarely see centers ever giving up sacks even though a few dominant DT's like Suh and McCoy are racking up sack numbers. I assume the sim engine lined them up against the tackle for all their sacks.

Based on what I can tell though, I don't think 10-15 year degradation is going to be *that* bad because from what I can tell in testing the drops occur mostly to the top 4-6 tackles that give up the most sacks, and as long as the draft brings in about 4-6 tackles that are decent pass blockers every year then it shouldn't fall apart into a land where there's nothing but 50-60 OVR tackles.

Agree
I need to test more though because I didn't check to see if the cpu teams still start those tackles. If they're benchdc/released/recycled I'm good with that. If they're playing and I can't fix that...no bueno.

Sent from my SCH-I545 using Tapatalk

[Bootzilla]

Re: Post patch LT regression

It's frustrating that last year the sim engine couldn't produce any sacks....so to offset that....this year they jack up the sim stats for sacks but unevenly distribute it against the tackles. I rarely see centers ever giving up sacks even though a few dominant DT's like Suh and McCoy are racking up sack numbers. I assume the sim engine lined them up against the tackle for all their sacks.

Based on what I can tell though, I don't think 10-15 year degradation is going to be *that* bad because from what I can tell in testing the drops occur mostly to the top 4-6 tackles that give up the most sacks, and as long as the draft brings in about 4-6 tackles that are decent pass blockers every year then it shouldn't fall apart into a land where there's nothing but 50-60 OVR tackles.

Honestly it doesn't look too bad. In an offline CFM I played the first game and simmed the rest. Had progression at the end of the season and only two LT's had poor pass block ratings and they were in the 60's. I believe 62. And both of those guys confidence was below 50. Everyone else, even those who gave up a decent amount of sacks, as long as their confidence didn't go below 50 they all remained the same or were higher in pass block. I honestly don't see this affecting but a couple of players per year.

[EricFreakingBerry]

Re: Post patch LT regression

Agree
I need to test more though because I didn't check to see if the cpu teams still start those tackles. If they're benchdc/released/recycled I'm good with that. If they're playing and I can't fix that...no bueno.

Sent from my SCH-I545 using Tapatalk

I tested some and some guys were still the best OVR option for their team even with 60-70 PBK but others were cut or relegated to backup duty. If they stay in at least it minimizes how many T's get decimated, but the downside to that is they may give up even more sacks until they've been effectively chased out of the league.

Sent from my LG-H811 using Tapatalk

[EricFreakingBerry]

Re: Post patch LT regression

Honestly it doesn't look too bad. In an offline CFM I played the first game and simmed the rest. Had progression at the end of the season and only two LT's had poor pass block ratings and they were in the 60's. I believe 62. And both of those guys confidence was below 50. Everyone else, even those who gave up a decent amount of sacks, as long as their confidence didn't go below 50 they all remained the same or were higher in pass block. I honestly don't see this affecting but a couple of players per year.

My test I had 4 LT's (Lewan, Oher, G. Robinson, and Matt Kalil killed to sub 70 PBK, plus 2 or 3 RT's who gave up 10-20 sacks too) But like I said most stay stable or even progress and we should always equalize the league with 5-6 decent rookies.

Sent from my LG-H811 using Tapatalk