Warrior General Games

After a brief discussion with Ihouse over PMs, we agree that this might well and truly just be a side-effect to the open-field pathing bug, and if that problem is solved, this one will be too. If someone can find a replay where a defender changes his interception course whilst a player is just running straight ahead (im not having any luck with this), we should continue discussing this. If not, then wouldn't it be best to just sort out the open-field pathing?

Originally posted by NiborRis
No, I don't think that's how it works. What happens is that the Defender(D) is mapping out an intercept course, and as the ball carrier changes his path from moving right to moving left, D is seeing a huge swing in intercept point, as he's not going to catch him anytime soon.
On top of that, I think vision checks are used to introduce errors in intercept point, and when the intercept point is very far away, small vision mistakes can add up to a BIG miscalculation in attack angle. If that's true, then it might be an improvement to change the vision checks to adjust the angle of the turn by small amounts rather than move the intercept point, but that doesn't work as well for things like attempting to intercept/deflect a pass, I think.

I'm not *overly* worried about it for the MLB and the CB, because in this case the TE is going faster than they are and they just can't catch him. Perhaps better logic could be put into place for pursuit from behind when the defender won't catch the ball carrier before the endzone - something like just moving directly for the ball carrier might look better - but that's mostly just going to try to make it look smoother. In fact, all the pursuit here is slower than the TE, so we're probably in "pursuit code does odd things when you won't catch the ball carrier" type of bug, but really bad movement behavior is still a bug, I think, and should be improved, even if it won't impact the result of the play.



just had another thought. Maybe it is adjusting to an attack point....but the angle is ALL messed up. Is the attack point bound within the confines of the field (in bounds)? (only Bort can answer that question). If it isn't, that might be causing at least part of the zigzag

Originally posted by kuaggie
just had another thought. Maybe it is adjusting to an attack point....but the angle is ALL messed up. Is the attack point bound within the confines of the field (in bounds)? (only Bort can answer that question). If it isn't, that might be causing at least part of the zigzag


Given the way the math would work, and the behavior of dot pursuit I've seen over the years...I strongly doubt it's bound by the field. Since we're talking about a dot that can't catch up here, I have no idea what the behavior drops out to, as there shouldn't be an intercept point defined. I think in the long distant path defenders even gave up and stopped if they couldn't catch a dot, right? (those replays are all long gone off the servers so I can't even go check) There's some code somewhere to keep them chasing, but it needs work.

I don't think is an actual bug in code as it is maybe an oversight in the way the pursuit code works. the SS seems like he is following every little cut the TE makes. If the SS isn't as fast, and he sees that, then, mimicking a human being would say that he should run straight at him, regardless if he's cutting or not.

I don't think its a bug, but i think its something that is overlooked in the code and should be fixed. is that a bug in GLB then? i'm not sure.

Just curious what others think of the logic of this, more from a football logic point of view than a GLB coding (since we don't know how this is coded and I don't want to assume).

Basically for a defender in chase mode (ballcarrier is closer to goalline than defender on Y axis) AND if the ballcarrier is faster than me, I'm just trying to get "close" to the play in case a teammate slows him down with a missed tackle, etc.

So, if ballcarrier is inside the "cone" drawn between me and the two pylons, I take a direct route from me to ballcarrier (except staying clear of teammates and blockers within reason). That means my vector may change slightly if ballcarrier weaves, but not dramatically (until/unless I get really close to him and he jukes... basically close enough that I'm going for a tackle, which in this case likely only happens if someone else slows him up).

If ballcarrier is outside of that cone (closer to sideline than if I ran straight towards pylon, or just short of it say the 5 yard line sideline), then I head straight towards that point instead. Not 100% sure on this, may be instances where a sharper angle is preferred, but generally speaking you're not going to catch up unless something happens (like missed tackle), so heading towards pylon gives most amount of time for something to happen for you to get back in the play. If/when a missed tackle slows carrier down enough that my Y axis (yardline) gets between him and goalline, then I look at (softly) adjusting my angle more towards ballcarrier (maybe somewhere in between pylon and ballcarrier), but nothing too abrupt so that I'd lose speed/momentum, and less abrupt the farther away I am along X axis from him. If I'm close to him, I'm going more aggressive to go for the tackle. If I'm still 10+ yards away (horizontally) then I'm still heading upfield of him a bit to try and get myself between him and the goalline and cut off the angle where he could accellerate past me again.

Otherwise, if I'm faster than ballcarrier, or if the ballcarrier is not upfield of me (I'm not in "chase" mode) then I use a more difficult and subjective logic of trying to intercept him while minimizing his yards gained. In reality (football logic) I'm also going to take into account the presence of any teammates who look likely to force him back inside, outside, etc.

So... I'm curious what folks think of all that. Tying it back to GLB, my hunch is that the current code is more along the lines of the latter more subjective "intercept" logic in most cases, whereas in reality the "football logic" when the player is upfield of you and faster than you is actually much simpler... just run to try and get back into the play, not directly to try and "make the tackle" necessarily. At that point you kinda just pick an angle and go with it until the situation changes, or you start sucking wind and give up!

If we're going for football logic, I think we need to make it even simpler:

1. If I think I'll never be able to chase him down, even if a team mate slows him down, I wont.
2. If I think I'll be able to close him down, I'll run ahead of him and try to cut him off, or push him to the sidelines
3. If I think I can stop him, I'll run just a bit ahead of him to make that tackle intercept and stop him.

Originally posted by NiborRis
Given the way the math would work, and the behavior of dot pursuit I've seen over the years...I strongly doubt it's bound by the field. Since we're talking about a dot that can't catch up here, I have no idea what the behavior drops out to, as there shouldn't be an intercept point defined. I think in the long distant path defenders even gave up and stopped if they couldn't catch a dot, right? (those replays are all long gone off the servers so I can't even go check) There's some code somewhere to keep them chasing, but it needs work.

my thoughts too.....and bounding it could potentially get rid of some of the swerviness

Originally posted by Djmr
If we're going for football logic, I think we need to make it even simpler:

1. If I think I cant chase him down, I wont.

No, at least not so simply... I've seen plenty of instances in both real football as well as GLB where a player slower than the ballcarrier during chase ends up making the play, because some teammate slowed them down. Regular occurrence.

It's only in the case of "there's no chance in hell I could stay close enough to make a play even if he stumbled and fell on his face" that you wouldn't chase... either that or if you have some sort of a thing for being berated by your coaches.

Originally posted by Stray Doug
It's only in the case of "there's no chance in hell I could stay close enough to make a play even if he stumbled and fell on his face" that you wouldn't chase... either that or if you have some sort of a thing for being berated by your coaches.

Oh, thats what I meant, editing original post then

Im beginning to think that this issue is becoming less and less of a bug. With this:

Originally posted by Djmr
After a brief discussion with Ihouse over PMs, we agree that this might well and truly just be a side-effect to the open-field pathing bug, and if that problem is solved, this one will be too. If someone can find a replay where a defender changes his interception course whilst a player is just running straight ahead (im not having any luck with this), we should continue discussing this. If not, then wouldn't it be best to just sort out the open-field pathing?

And this:

Originally posted by Catch22
Not necessarily a bug if it doesn't mirror football logic. Sometimes GLB has to do things that don't mirror football logic (like having the RG pull to the left side to pick up the blitzing ROLB) to make the game work. You need to evaluate if it's causing issues with game balance - if it is, then it's probably a bug. If it's not, then it's probably something we'd want to look at as an enhancement/suggestion to try and fix.

It leaves two questions on me:
1. Is it purely an open-pathing thing?
2. Does it not affect game balance?

If yes to either one, we should mark it down as an issue determined not a bug. If no to both, its best to file it. Anyone agree?

Originally posted by Djmr
It leaves two questions on me:
1. Is it purely an open-pathing thing?
2. Does it affect game balance?

If yes to either one, we should mark it down as an issue determined not a bug. If no to both, its best to file it. Anyone agree?

(Side note - your last line is backwards, but I know what you meant)

My feelings:
1) Maybe. It's possible that we'll see similar bad pursuit effects from a legit cut - a CB or LB chasing the play from behind while an elusive back makes a move to cut around/between/etc the safeties, for example. Chasing a player from behind who is making small cuts shouldn't cause the pursuit to make huge cuts.

2) It leads to breakaways, so I think so. The example we have in hand has the pursuit slower than the ball carrier - maybe we should try to get some examples where the pursuit is faster? It means plays are much harder to run down from behind, which affects how Defenses will try to defend.

Edited it to make sense

I feel the same way, looks like we may need more information here. Im going to wait and see what the others think.

I think #1 is unknown for certain, but it seems to be the primary case. I don't know how you could definitely say it is "not" more than an open-pathing thing, like, you'd have to analyze every possible non-open-pathing angle/situation in order to "prove" it's purely open-pathing. I think that's an unreasonable expectation. But yes, it does seem primarily an open-field pathing issue, and we've seen no evidence to the contrary.

For #2 I'd say "maybe slightly" but don't think it's affecting game balance in a significant way. I guess part of the question here is how we are defining the issue. If we define it as open-field pathing only in the case where D is slower than ballcarrier (which I think we should) then I think this is a suggestion for improvement and not significantly affecting game balance, it's more just an annoyance that should be improved. The replay in the OP shows no problem for a player faster than ballcarrier (FS), so imo let's not conflate that situation into the issue that was posted here (which appears to be only players who are [b[slower than ballcarrier chasing in an erratic manner).

Also, this is potentially related to bug #159, although that bug appears to be for the "offensive" player's problem (open field cuts to avoid chasing defenders are too exaggerated/dramatic and causing loss of speed). That's clearly a bug as it causes player on O to be tackled when they should be breaking away.

The bug here is that players who "most likely aren't going to make a play anyway" are weaving too much. *shrug* I don't think it rises to the same level, doesn't have same impact on the game. I do think it's something that should be improved. My sense is that if/when Bort looks at bug #159 for the offensive side, he should also take a look at this component of the issue as well just to be comprehensive about it, and see if there's a way to clean it up a bit.

we all seem to be of similar opinion. It looks weird, but how it would affect the game is fairly minimal. I like stray doug's idea of incorporating this into bug #159.....and then marking this as not a bug or needs QA admin determination (just to consolidate things)

Different type of play, but another example of bad pathing..... just so it doesn't get forgotten.

http://goallineblitz.com/game/forum_thread.pl?thread_id=4077787

fwiw, i really don't think that one is a bug, MC.