Warrior General Games

I'm going to honestly try to make this at least a weekly thing. If there isn't a new one posted by June 24th, PM me and bug me about it

Only 4 Major Topics this time, but 3 of them are pretty huge topics, and are pretty much getting all of our attention.

░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░Major Topics and Additions░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░

#1 - Biggest Topic is obviously talking about the DE/OT interaction after the changes that Bort put in after the offseason. There's a fairly general consensus that we're at least CLOSE to where we want to be. Some think it's too low, some think it's too high, but we all think it's close.

#2 - Sometime this season the number of scrimmage slots is going to double, as Bort moves the scrimmages to their own dedicated server. Also talking about some other options to help the major tournament coordinators plan their stuff.

#3 - I'm working with the Offensive Coordinators forum to try to get some new plays added into the playbook. I had them all make some plays and submit them, and I'm in the process of organizing it all now, then we'll start some discussion on which ones to use. Not sure on the timeframe of when these will be put in, just know that we're working on it.

#4 - A LOT of talking about the major upcoming project, being able to turn retiring players into coaches. Mostly based around these two different suggestions:
http://goallineblitz.com/game/forum_thread.pl?thread_id=1423086
http://goallineblitz.com/game/forum_thread.pl?thread_id=293787

░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░Small Bugs Being Looked At░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░

└Blitzers not picking up on handoffs or QB runs (general consensus is that it is failed vision checks)
└High level CPU players in low-level capped leagues
└Players running out of the back of the endzone
└Coverage on high hang time/short punts
└FB blocking logic a little off, especially for fast FBs

░░░░░░░░░░░░░░░░░░░░Quick Hits - Topics and Suggestions that have been briefly mentioned.░░░░░░░░░░░░░░░░░░░░
└Feasibility of adding injuries to the game
└Adding timestamps to test game list
└Hang Time as a stat for punters
└A way to share and rank DPC plays
└AI Test link direct from PBP

As always, you can see the actual test games being run here:
http://test.goallineblitz.com/game/game_list.pl

interesting - thanks!

i don't want to be too negative - but i hope the 'coaching' project gets extremely and thoroughly tested...and run through the ringer before going live -

i don't think VA's were polished enough when they were introduced (maybe not even now!)

***************

r8 - are you - or any of the testers interested in doing an extensive SA test? I'm not even convinced that 75% of them even work -

It would be a huge job - but i'd really like some more definitive confirmation that they do anything - and more information would only help (as long as it's published) -

things like: how does an O-Line - all with 10 in run block compare to an O-line with no SA's (and exclusively cut block; exclusively shock block; etc etc for every position).

I think there are some major issues there - and though Bort has been giving a bit more information lately - I don't think it's coming quick enough.

Also - with the ability to choose an offensive play (maybe you always had that?) - it might make the testing/analysis more doable than it was in the past

Originally posted by TxSteve
interesting - thanks!

i don't want to be too negative - but i hope the 'coaching' project gets extremely and thoroughly tested...and run through the ringer before going live -



Thanks for doing this every week, r8. But I also wanted to second the concern that the coach thing be really well hashed out. In particular, I think it would be nice if some of the admins thinking about Coaching were shared with the user base, even before they were finalized. It's good they're talking about it with the testers, but it would be even better if they got everyone involved.

Originally posted by TxSteve
interesting - thanks!

i don't want to be too negative - but i hope the 'coaching' project gets extremely and thoroughly tested...and run through the ringer before going live -

Well, even before the test server came up I had been talking with Norse thinking through the whole coaching thing, probably starting back sometime around Season 4 or 5, so there's been a LOT of talk about it. This isn't the first time it's come up on the test server, either, just now it's time to really start trying to work out and finalize things because it needs to be ready to go pretty soon.

And yeah (@ Capaneus), that's part of the reason I posted those two links, Bort WANTS feedback from the community on this one. Those two links are the two different ways that are being considered, so if you want to chime in, head to those topics and put in your $0.02.

Originally posted by TxSteve
r8 - are you - or any of the testers interested in doing an extensive SA test? I'm not even convinced that 75% of them even work -

It would be a huge job - but i'd really like some more definitive confirmation that they do anything - and more information would only help (as long as it's published) -

things like: how does an O-Line - all with 10 in run block compare to an O-line with no SA's (and exclusively cut block; exclusively shock block; etc etc for every position).

We're still really trying to avoid this kind of stuff. If there's something that DEFINITELY has a problem with it then we'll look at it, but I don't think there's really anything that seriously wrong with any of them.

That being said, we are still looking at changing Lower the Shoulder to Stiff Arm for TEs, and some other things get discussed every once in awhile, but as a whole we're trying to avoid those kind of things that would simply give out too much information about the game.

So, since it isn't mentioned in your post, I can assume that the drifting of linebackers will remain the attempt du jor to open up the outside runs, and short halfback passes?

And that the DEs are still slowing down on evasive pass rushing instead of acting like an actual DE and going balls to the walls to get around the tackle?

Stop nerfing the defenses to make up for bad blocking schemes.

PinTBC

Originally posted by PinTBC
So, since it isn't mentioned in your post, I can assume that the drifting of linebackers will remain the attempt du jor to open up the outside runs, and short halfback passes?

And that the DEs are still slowing down on evasive pass rushing instead of acting like an actual DE and going balls to the walls to get around the tackle?

Stop nerfing the defenses to make up for bad blocking schemes.

PinTBC

Ok now, I like having people around to disagree and play devil's advocate and all that, but seriously, your opinion on the drifting LBs has been duly noted

I know that you don't like the LB drifting, I know that you think DEs are still messed up, and I know that you don't like the blocking schemes. You really don't have to keep saying it in EVERY topic. I'm to the point that I don't even really need to read what you're saying anymore, because as soon as I see that you've posted in a topic, I know exactly what it says.

However, as I've mentioned 100 times before, there are no easy solutions to these problems, especially with the blocking scheme. We've discussed it time and time again and have never really been able to figure out a better method then what is being used right now. We're still trying to think of ways to improve it, but it's not exactly an easy task, and complaining about it in every topic isn't going to make the process any faster.

Now, if you have some brilliant idea to make it work better, please do share, but until then, lay off a bit ok?

R8 - with respect, I think the LB drift thing has been given short-shrift as a bug that needs addressing. Bort said it was fixed - it isn't, its affecting outcomes, and just because you're tired of PinTBC mentioning it all that time doesn't mean that it isn't an issue worthy of more attention than it *seems* to be getting.

Originally posted by OttawaShane
R8 - with respect, I think the LB drift thing has been given short-shrift as a bug that needs addressing. Bort said it was fixed - it isn't, its affecting outcomes, and just because you're tired of PinTBC mentioning it all that time doesn't mean that it isn't an issue worthy of more attention than it *seems* to be getting.

We know of the problem, but can't really figure out a way to fix it.

For one, it isn't happening very often anymore, so it's not exactly easy to track down.
For two, part of it is somewhat built-in because LBs aren't supposed to be omniscient anymore, they're supposed to take a second to read the play.

And besides, if it was really that big of a problem, why aren't more people complaining about it? Besides Pin saying it in damn near every topic I make, I really don't see it coming from anyone else as a major problem.

Originally posted by r8


And besides, if it was really that big of a problem, why aren't more people complaining about it? Besides Pin saying it in damn near every topic I make, I really don't see it coming from anyone else as a major problem.

Do you know why Pin mentions it whenever he can?

It's because ever since you posted that it was supposedly "fixed" I haven't seen, or read a single reply to any of my posts about it. I've been getting a glaringly loud silence. It didn't even make any of your list of problems being addressed, so why in hell would I even suspect someone might be looking at it?

As for your DE comments, yes, I firmly believe Bort is screwing up by addressing Defensive line problems by modifying the quarterbacks. It might manage to gain you a 7% sack rate in the Pro levels, but you'd be hard pressed to make a statement and claim that it actually SHOULD get you that 7% sack rate there. That 7% should be when an average offensive lineman, goes up against an average DE, better/worse builds should differ greatly. With the extra allowances that have been programmed into the game to give bonuses to offensive linemen to allow them to get outside and block, they move much faster on average than they should which screws up the balance in itself.

As for your blocking comments, even you have to agree that there re some serious problems with the blocking and blocking assignments currently in the Sim. Hell, there are whole threads about FBs and TEs ignoring what should be their main targets. The reason I keep bringing that one up is that if Bort ever does manage to even and clean up his blocking schemes, you are going to once again seriously unbalance the offensive and defensive interactions. I've been an engineer for a very long time now, and I understand very clearly that if you don't fix the basic building blocks of your system (any system), you are fucked in the long run.

Originally posted by r8
For one, it isn't happening very often anymore, so it's not exactly easy to track down.

It actually happens quite often. If Bort is correct, it happens after the linebacker makes his vision check. The best way to test for what is going on, is for Bort to run his own sim with all his numbers shown, and step through the sim. Actually the better way to test for it would be to put the numbers on the tester sims and let them run a few games, it would certainly rear its head and should be much easier to chase. For whatever reason Bort doesn't want to test that way, so he is stuck doing it himself.

Originally posted by r8
For two, part of it is somewhat built-in because LBs aren't supposed to be omniscient anymore, they're supposed to take a second to read the play.

Hence the vision check they have to make before they can even move. I have no problems with that other than if they have to make vision checks to do that then blockers need to be forced into vision checks to pick up bliltzers, and blocks around the end. The problem is slowing them down after that vision check (according to Bort).

As much as you get sick of reading my posts, I really hate simply talking to a wall.

Originally posted by R8
And besides, if it was really that big of a problem, why aren't more people complaining about it? Besides Pin saying it in damn near every topic I make, I really don't see it coming from anyone else as a major problem.

Oddly I've asked myself the same question, and even gone out and watched other people's games to see if maybe something is just fucked up with my training. I've seen the linebackers do this on every single sim I've ever looked at. I think people have just given up and decided to push speed again, which can make up for the slow drift, as once the linebacker does start moving he needs more speed now than ever to run down the target.

PinTBC

good stuff, R8. Nice to see the lowly punters getting some minor attention.

Originally posted by r8
Now, if you have some brilliant idea to make it work better, please do share, but until then, lay off a bit ok?

Fair enough, I'm in a pissy mood right now.

If I were the designer of this game, I would use what is effectively a zone blocking scheme. Without drawing this, it will be difficult to explain, but here goes.

Each blocker (key word blocker) at each click has a range of notice, an area of effect, a range of engagement, and a threat value for targets.

When the offense is rushing, the offensive line sets up and evaluates the targets in their range of notice (distance from dot), and area of effect (direction of facing). The threat value is assigned by planned movement lane, target motion (blitzers higher priority), and distance from dot. Each threat value has a vision component because a lineman can simply choose the wrong guy to block. Each click the threat value is updated as the blocker moves along his path, and when the target arrives in his range of engagement (he can get his hands on the target) he blocks/grabs/engages him.

The key is that each click you need to have the blocker reassess the potential threats in the rushing lane.

On a passing play, the blockers start out with the same assessment, but instead of moving forward, they are falling back into their pocket. Each click they assess the threat modified by vision (I'd add in a history component here also if a NT keeps beating the center, a Guard may decide to help out). If a man is blocked, their threat drops significantly unless they start to beat their man. This allows Guards to keep their heads up and look for blitzers instead of automatically doubling and tripling down on the NT. The DE and tackle interaction would end up in an attribute race between the speed of the DE versus the agility of the tackle, along with the strength of the DE versus the strength of the Tackle, hopefully you will eventually get the option to create offensive play sets like the DPC and then you can adjust your tackles to handle split DEs etc..

Now by the way, I would change the way the defense gets to rush also. They simply rush too fast from the get go. They should be making vision checks as the plays develop to determine which hle they will attack, or which direction the play is going, but when they make that check, they should ramp up to top speed using a combination of vision and agility. By the way, the defense should be able to read the play wrong and hit the wrong hole trying to stop it. Say the outside linebacker 'reads' a guard tackle hole and hits it, while the HB decides to bounce it outside. That should happen every once in a while.

I realize that what I've described isn't far from what we have, and that is what is so frustrating. Bort started with a pretty damned decent game all the way back in season 3. It needed some cleaning up of the interactions specifically along the offensive and defensive lines, but he ended up starting down the road of buffing and debuffing actions, which actually make it even harder to balance in the long run.

But whatever. I need another beer.

PinTBC

As nitpicky as some posts can get, your above post seems reasonable as an ex-college player. Our offensive line used zone-blocking schemes almost exclusively unless there was some super stud on the other team. Each guy kept his 'head on a swivel' looking for potential shenanigans.

In addition, mostly the C would make line calls, possibly further increasing the need for solid vision/line general attributes for the C instead of just being a STR/BLK whore to adjust from zone blocking to double or triple teaming monsters on the DL.

I approve.

Originally posted by PinTBC
If I were the designer of this game, I would use what is effectively a zone blocking scheme. Without drawing this, it will be difficult to explain, but here goes.

Each blocker (key word blocker) at each click has a range of notice, an area of effect, a range of engagement, and a threat value for targets.

When the offense is rushing, the offensive line sets up and evaluates the targets in their range of notice (distance from dot), and area of effect (direction of facing). The threat value is assigned by planned movement lane, target motion (blitzers higher priority), and distance from dot. Each threat value has a vision component because a lineman can simply choose the wrong guy to block. Each click the threat value is updated as the blocker moves along his path, and when the target arrives in his range of engagement (he can get his hands on the target) he blocks/grabs/engages him.

The key is that each click you need to have the blocker reassess the potential threats in the rushing lane.

On a passing play, the blockers start out with the same assessment, but instead of moving forward, they are falling back into their pocket. Each click they assess the threat modified by vision (I'd add in a history component here also if a NT keeps beating the center, a Guard may decide to help out). If a man is blocked, their threat drops significantly unless they start to beat their man. This allows Guards to keep their heads up and look for blitzers instead of automatically doubling and tripling down on the NT. The DE and tackle interaction would end up in an attribute race between the speed of the DE versus the agility of the tackle, along with the strength of the DE versus the strength of the Tackle, hopefully you will eventually get the option to create offensive play sets like the DPC and then you can adjust your tackles to handle split DEs etc..

Ready to have your mind blown?

That's basically already how it works, minus the "direction facing" component. Which is something that I've told you at least 5 or 6 times that I think would fix the problem, but would require a major overhaul of just about the entire system.

O-Lineman do reassess the risk on every tick, they do determine who they think the biggest risk is, and there is a significant drop in the risk level if someone is already blocked.

Now do you see our problem?

Originally posted by PinTBC

Now by the way, I would change the way the defense gets to rush also. They simply rush too fast from the get go. They should be making vision checks as the plays develop to determine which hle they will attack, or which direction the play is going, but when they make that check, they should ramp up to top speed using a combination of vision and agility. By the way, the defense should be able to read the play wrong and hit the wrong hole trying to stop it. Say the outside linebacker 'reads' a guard tackle hole and hits it, while the HB decides to bounce it outside. That should happen every once in a while.

I realize that what I've described isn't far from what we have, and that is what is so frustrating. Bort started with a pretty damned decent game all the way back in season 3. It needed some cleaning up of the interactions specifically along the offensive and defensive lines, but he ended up starting down the road of buffing and debuffing actions, which actually make it even harder to balance in the long run.

Ok, now I see that you understand that it's basically how it works already, so I'm not real sure why you even bothered with all of that Now yes, what you're saying about the D line being able to see and choose which hole they want to attack is a neat idea, but with the blocking problems that we're already having, I can't really see it creating anything more then more trouble. Maybe if we implement the "direction facing" stuff sometime then it could work, but with the current state of the sim... I highly doubt it.

And the whole buffing and debuffing thing is called adjustments. There hasn't been a single thing that Bort has made or put into the game that didn't require some adjusting. What's making it harder to balance isn't because he decided to make some adjustments, but rather the growing number of variables in the system as he tries to make the sim more complete.

Originally posted by r8
Ok, now I see that you understand that it's basically how it works already, so I'm not real sure why you even bothered with all of that Now yes, what you're saying about the D line being able to see and choose which hole they want to attack is a neat idea, but with the blocking problems that we're already having, I can't really see it creating anything more then more trouble. Maybe if we implement the "direction facing" stuff sometime then it could work, but with the current state of the sim... I highly doubt it.

So, if his basic structure is there (like I said it has been there since season 3 at least), then debug it and figure out why it isn't working as well as it should. Yes, I know this is difficult, so is debugging a system simulation that predicts the effectiveness of a missile defense on an aircraft. The vision component is a big thing, but his decision logic has some flaws in it. The blockers often select a target early in a play, and instead of making a decision every tick, they remain focused on their selection throughout the play. That's what was making the delayed blitzes by linebackers so effective. Fix that decision process, and you no longer have to slow linebackers down, you have to work the threat and engagement logic of the blockers versus the block avoidance of the linebacker.

Same thing for a fullback coming through the guard/tackle gap. He looks for his target all the way from the start of the play, and once he is through the gap, or finds someone in the gap you end with a block or avoid block check. This ideaof giving fullbacks bursts of speed so they can get in front of their halfbacks is wrong. If you want a fullback to lead block, then they damn well better be almost as fast as your halfback, or he should delay some and follow. That is the same situation with these sweeps. If you want a pulling guard to lead your halfback around the corner, then you better design your guard with some significant speed, and agility or he will never get there. Bort shouldn't have to design a boost to allow blockers not to need speed/agility attributes.

Originally posted by r8
And the whole buffing and debuffing thing is called adjustments. There hasn't been a single thing that Bort has made or put into the game that didn't require some adjusting. What's making it harder to balance isn't because he decided to make some adjustments, but rather the growing number of variables in the system as he tries to make the sim more complete.

It actually is a huge part of his problem. He has been buffing and debuffing based on current player builds, and what that eventually forces is exactly one build that works. We've been seeing that on DEs for a couple seasons now, where the only build that semi works is a hugely high agility and speed. The builds Bort mentioned that use "stupid" agility.

Buffs and debuffs only allow players to make up for a lack in an area of the attribute training. I use the term debuff as an actual lessening of an attribute to attempt to balance an equation. I was actually on Ken1's side when he was complaining about how effective speed DEs were back in season 7, because they were so fast they were blowing through a mechanism that should have been there. The blocking mechanism was pretty much hosed because they never triggered it. Much of that was due to tackles not being agile or fast, and Bort's changes ended up being a buff of a tackle's ability to get to a position, and a bebuff of the speed a DE could use. Since the tackles didn't need high speed, or agility many of them are pumping strength and blocking through the roof, which means once they make contact, the DE is simply dead. That is an example of why buffing and debuffing before the basic mechanism works well is simply a piss poor way to "fix" a complex system, and yes, even though this is a fairly simple game, the interactions between the components makes it a fairly compex system which by its nature is difficult to debug.

That is why I've been so dead set against the test server, and the testing of these interactions with little or no insight into the actual numbers of the game. I'm actually beginning to believ ethe reason Bort refuses to put the numbers on screen so he can actually debug some of this stuff is because the dots have nothing to do with the play. He figures out what happens in the background using a simple statistical equation, and simply has the dots run around to kind of show the player what happened. That would actually explain a significant amount.

Thank you for at least commenting on my previous rant.

PinTBC

With respect to coverage on short, high punts, what is being looked at? I guess this thread is outdated but it wasn't posted as a bug in the other thread.