New Sticky Note

NOTICE: This league is using the BLEEDING EDGE game engine. For more information, click here.

League Forums

Main - Beta Chat

0.4.5 Observations

By setherick
2/05/2019 8:11 pm
1) WR Courage actually needs to be boosted now that ***** are being thrown closer to them. Not boosted by much, but the algorithm that compares Courage to Punish needs to be rethought. There are too many knockdowns now oddly enough.

2) It's almost impossible to throw to the middle of the field because of how defenders react to the ball. Defender reaction time should probably be driven off an attribute like Intelligence. The smarter the defender, the better they are at getting to the ball.

3) All of the catching nerfs that are still in place need to be removed. Basically any receiver with better than 50 catching should almost never drop a pass if he's wide open.

Re: 0.4.5 Observations

By setherick
2/06/2019 7:22 am
setherick wrote:
3) All of the catching nerfs that are still in place need to be removed. Basically any receiver with better than 50 catching should almost never drop a pass if he's wide open.


Here are a bunch of examples of what I mean:

1) https://mfn1.myfootballnow.com/watch/9318#1697126
2) https://mfn1.myfootballnow.com/watch/9318#1697222
3) https://mfn1.myfootballnow.com/watch/9318#1697228
4) https://mfn1.myfootballnow.com/watch/9318#1697257
5) https://mfn1.myfootballnow.com/watch/9318#1697271

Nerfing catching never made sense, but now it definitely doesn't make sense since DBs actually cover their man.

Re: 0.4.5 Observations

By jdavidbakr - Site Admin
2/06/2019 9:32 am
The only 'nerf' that really still exists is a decrease in the catch probability for longer passes, which I feel like has to remain (longer passes are less likely to be completed in real life) or else we'll be back to the all-long-pass game plans that are unstoppable. The catch algorithm essentially starts with the WR's catch ability, then it is reduced based on punish vs courage (which takes player heights and velocity vectors into account) and pass accuracy (which has been calculated based on the accuracy attribute, pressure, scramble skill, and distance - which has less impact with higher strength). That's pretty much all that decides whether the ball will be caught or not. A case can be made to tweak the weights that each of those manage, but I am also getting 60-70% completion rates with the testing server which is where the NFL goal is. So to increase completion probability we also have to decrease it somewhere else or we're going to have too much strength in the passing game.

If a WR is wide open, he may drop the ball because the QB was under pressure or didn't have the skill to throw the ball well, but there is also a non-zero probability that any ball will be dropped no matter how perfectly it is thrown.

Re: 0.4.5 Observations

By raymattison21
2/06/2019 11:16 am
tweaking or rewriting End route senerios , qb scrambling /throwaway , qb reads, knockdowns and INTS in that order Wil help numbers

Scenarios are lacking for me to compare to nfl numbers much more . Nfl plays more zone too. This playoffs has drop numbers lower than the nfl.

Visually that up for debate with our schemes and talent I don't see many bombs attempts . Opening up (lowering) deep penalties is up for debate . IRL dbs have more recovery time and weak armed guys can t make throws . There no penalty it's harder do to physics. Maybe your is that good .

Re: 0.4.5 Observations

By setherick
2/06/2019 9:38 pm
So there is - I am now convince - a big bug with the accuracy calculation that is leading to a lot of interceptions. And a fix for it should be pushed to all leagues.

Here's what I'm almost certain is happening. The accuracy calculation consider how open a WR is compared to the defender that is guarding him. It DOES NOT, again almost certain, take into account defenders in the vicinity.

The reason this matters is because when the defenders were changed to react better to the ball in the air that made nearby defenders equally strong defenders since QBs loop a lot of passes. This means QBs should NOT be throwing to WRs where there are multiple nearby defenders that can make a play on the ball.

If I'm right, and I'm pretty sure I am, this needs to be a hot fix.

Re: 0.4.5 Observations

By jdavidbakr - Site Admin
2/07/2019 11:30 am
setherick wrote:
So there is - I am now convince - a big bug with the accuracy calculation that is leading to a lot of interceptions. And a fix for it should be pushed to all leagues.

Here's what I'm almost certain is happening. The accuracy calculation consider how open a WR is compared to the defender that is guarding him. It DOES NOT, again almost certain, take into account defenders in the vicinity.

The reason this matters is because when the defenders were changed to react better to the ball in the air that made nearby defenders equally strong defenders since QBs loop a lot of passes. This means QBs should NOT be throwing to WRs where there are multiple nearby defenders that can make a play on the ball.

If I'm right, and I'm pretty sure I am, this needs to be a hot fix.


The openness of a receiver does take into consideration where the WR is running and the defenders in that area. It actually does not care at all about whether a defender is actually assigned to cover the WR.

Re: 0.4.5 Observations

By setherick
2/07/2019 5:52 pm
jdavidbakr wrote:
setherick wrote:
So there is - I am now convince - a big bug with the accuracy calculation that is leading to a lot of interceptions. And a fix for it should be pushed to all leagues.

Here's what I'm almost certain is happening. The accuracy calculation consider how open a WR is compared to the defender that is guarding him. It DOES NOT, again almost certain, take into account defenders in the vicinity.

The reason this matters is because when the defenders were changed to react better to the ball in the air that made nearby defenders equally strong defenders since QBs loop a lot of passes. This means QBs should NOT be throwing to WRs where there are multiple nearby defenders that can make a play on the ball.

If I'm right, and I'm pretty sure I am, this needs to be a hot fix.


The openness of a receiver does take into consideration where the WR is running and the defenders in that area. It actually does not care at all about whether a defender is actually assigned to cover the WR.


Then one of these two things is happening:

1) Passing velocity is too slow, which is allowing four or five defenders to converge on every longish (>10 yards) pass.

or

2) The calculation is way off. Because four or five defenders can converge on every longish pass, which should reduce the probability that the QB mentally calculates to drop significantly.

Basically, I'm trying to figure out why you can't throw the ball longer than 10 yards ever. I know one reason is because of how QBs "lead" WRs on long routes.

Re: 0.4.5 Observations

By setherick
2/08/2019 7:10 am
jdavidbakr wrote:
The only 'nerf' that really still exists is a decrease in the catch probability for longer passes, which I feel like has to remain (longer passes are less likely to be completed in real life) or else we'll be back to the all-long-pass game plans that are unstoppable.


I disagree. I find the catch rate for long passes to be completely unacceptable. They are so rarely thrown anymore that it makes this look utterly ridiculous. Here are some examples from a recent private league game:

1) https://usflwfl.myfootballnow.com/watch/2667#492853 - WR fights, fights, fights, gets free and drops the ball. The WR has to work for three seconds to get open. Let him catch the **** ball.

2) https://usflwfl.myfootballnow.com/watch/2667#492881 - WR makes a great move to get free of coverage and the QB does a decent job of leading him. But since this pass is over 10 yards, it's a drop.

3) https://usflwfl.myfootballnow.com/watch/2667#492890 - WR fights to get free of his man and QB stays strong in the pocket. But it's a drop because of the "decrease" in probability.

The decrease in probability is an unrealistic nerf and it should be removed immediately as a hotfix. WRs have to do a lot more work to get open, and QBs are less likely to throw long passes anyway.

Re: 0.4.5 Observations

By raymattison21
2/08/2019 10:36 am
setherick wrote:
jdavidbakr wrote:
The only 'nerf' that really still exists is a decrease in the catch probability for longer passes, which I feel like has to remain (longer passes are less likely to be completed in real life) or else we'll be back to the all-long-pass game plans that are unstoppable.


I disagree. I find the catch rate for long passes to be completely unacceptable. They are so rarely thrown anymore that it makes this look utterly ridiculous. Here are some examples from a recent private league game:

1) https://usflwfl.myfootballnow.com/watch/2667#492853 - WR fights, fights, fights, gets free and drops the ball. The WR has to work for three seconds to get open. Let him catch the **** ball.

2) https://usflwfl.myfootballnow.com/watch/2667#492881 - WR makes a great move to get free of coverage and the QB does a decent job of leading him. But since this pass is over 10 yards, it's a drop.

3) https://usflwfl.myfootballnow.com/watch/2667#492890 - WR fights to get free of his man and QB stays strong in the pocket. But it's a drop because of the "decrease" in probability.

The decrease in probability is an unrealistic nerf and it should be removed immediately as a hotfix. WRs have to do a lot more work to get open, and QBs are less likely to throw long passes anyway.


They were all I the third quarter . I see more drops like this as the game passes . Perhaps the parabolic change didn't favor fatigue calculations for skill ratings in these senerios .

Or the knockdown /drop logic needs to be scrapped

Re: 0.4.5 Observations

By setherick
2/08/2019 7:40 pm
It has nothing to do with the 3Q. It has everything to do with the drop calculation on long passes that was added as part of 0.4.3. That code was illogical then. It's worse now.