Warrior General Games

Originally posted by Staz
Originally posted by odg62

I cant believe i forgot this one....

Stop with the running backs being able to go through a pile of 4-8 guys like they arent physical objects subject to the same laws of physics of everyone else and make them have to run around the 300 pound lineman right in front of them

On another point

- The bigger your player (weight and height) the more space he should take up. Dots can stay the same size, but there should be a "true size" within that dot. The larger the player, the bigger his "true size" is, and the more space he takes up. This would change the inside run game, blocking, getting through blocks, etc. towards a more realistic nature. If this is already the case, it sure doesn't seem like it.

+1

One of the easiest fixes to the passing game is simply to make QBs actually throw to open receivers immediately (especially when set to -100 open man) - at the moment the QB seems to spend so long on vision checks that he almost always leads into coverage. Agree that there are 100 other things to fix on the passing game, but this drives me crazy.

Originally posted by ryanshaw
One of the easiest fixes to the passing game is simply to make QBs actually throw to open receivers immediately (especially when set to -100 open man) - at the moment the QB seems to spend so long on vision checks that he almost always leads into coverage. Agree that there are 100 other things to fix on the passing game, but this drives me crazy.

While I agree the decision making needs work, one of the biggest things people need to realize is that real life QBs are not robots. They do not see every window or hit every open receiver.

QBs can't be perfect, both in the real world and the dotball world, or defending the pass will be an absolute nightmare.

I actually think that most of the decisions are correct. You have to take into account the builds that we are playing with. At WL most of the problems are not an issue because the builds are good enough.

It is part-and-parcel of the build structure we have that no positions acts properly and logically. You can't have a QB with 80 throwing and 10 vision and expect him to see anything.

Originally posted by Mat McBriar
While I agree the decision making needs work, one of the biggest things people need to realize is that real life QBs are not robots. They do not see every window or hit every open receiver.

QBs can't be perfect, both in the real world and the dotball world, or defending the pass will be an absolute nightmare.

i dont think anyone wants the QBs to be perfect- that would be crazy but completing less than 50% of passes and general not great decisions arent making things much fun either
look at this play http://goallineblitz.com/game/replay.pl?game_id=1192820&pbp_id=19288359

why does the QB pass up the 1 on 1 coverage to throw into a man covered by 4 guys - at least go to the receiver who has a chance to make a play

Check his build. if he has no vision or confidence, then the numbers told the sim he couldn't make the throw. Also, the QB may have judged the CB to be too good for that WR so didn't risk it. The throw in the end was a reaction to the guy getting the hurry.

Originally posted by CONN CHRIS
http://goallineblitz.com/game/forum_thread.pl?thread_id=3835684

One of several threads on the subject of the D AI inserting random un-called plays. The difference here is that it was with a team that had an ultra-simplistic 4 input D AI (all created plays). If there is a test where you should be able to identify why this happens, I'd assume that a scenario such as this would be close to it.

Have Bort or anyone else gotten any closer to figuring out why this happens and how to prevent it?

It appears that no one else is bothered by the fact that the AI inserts random un-called D plays into the sim and they have no idea why which is good because they have no idea how to fix it.

+ Spin is overpowered and needs to be reduced a little, rather than completely nerfed
+ The confidence morale issue needs to be balanced better between the O and D, especially for WRs
+ More plays when no-one is beaten, pancaked or revcaked on the lines. Less extreme results basically.
+ Passing difficulty should be kept in proportion to the level - level 1 teams should be able to pass on a level 1 defence as well as level 70 teams can on a level 70 defence
+ The AI needs to follow the inputs more exactly - there are too many occasions when the plays it calls are at the outsides of probablility. In particular the offensive AI calls more passes than you'd expect.

Originally posted by Xar
I actually think that most of the decisions are correct. You have to take into account the builds that we are playing with. At WL most of the problems are not an issue because the builds are good enough.

It is part-and-parcel of the build structure we have that no positions acts properly and logically. You can't have a QB with 80 throwing and 10 vision and expect him to see anything.

Which is exactly why the cap-building system needs to change somehow. Either tie more attributes together (the way speed and agility are) or redesign the player building system. You just can't have a game that's only balanced for ~10% of the players.

I just dont see how I can have two DT/NT with over 115 strength getting knocked down consistantly...Also he never seems to push the pocket even single teams...I dont expect him to get sacks just plug the hole and get the occasional hurry and leaving the sacks to the DE's and Blitzers

Originally posted by nmariner
I just dont see how I can have two DT/NT with over 115 strength getting knocked down consistantly...Also he never seems to push the pocket even single teams...I dont expect him to get sacks just plug the hole and get the occasional hurry and leaving the sacks to the DE's and Blitzers

115 is nice, but I have had Centers with 148 Strength, and other O Linemen not far behind.

Some games you are just going to be the Spatula, other days, its time to play the Pancake.

Originally posted by sam_acw
+ The AI needs to follow the inputs more exactly - there are too many occasions when the plays it calls are at the outsides of probablility. In particular the offensive AI calls more passes than you'd expect.

I am more concerned with the AI calling plays that are Flow Chart impossible to reach with the settings put in on the OAI or DAI.

Has happened more than once, verified by Mods.

Its frustrating to have that much time dedicated to making a game plan and then have it go balls up and cost you a TD or the game or the satisfaction that you at least DID something with all that time.

Originally posted by Xar
I actually think that most of the decisions are correct. You have to take into account the builds that we are playing with. At WL most of the problems are not an issue because the builds are good enough.

It is part-and-parcel of the build structure we have that no positions acts properly and logically. You can't have a QB with 80 throwing and 10 vision and expect him to see anything.

Well no.

But then again if the QB was that horrible he would not actually be in the league, would he?

IMO the game needs to "Curve" the system a bit more. Its unbalanced and unsatisfying to have teams get rolled 255-0 or 100-0 or even 75-0 on a regular basis. Its also stupid. What league exists like this?

I think the game needs to have an underlying presumption that if a player is in a league that that player is competitive enough to play in that league, and to adjust his playing abilities accordingly. Consider it Playing Up - and Playing Down - to the level of competition if you wish to rationalize it beyond that.

Basically take all the players in a league and average their effective Level (by position if you like or by side of the ball or just over all) and then "curve" all players abilities so that they are no more than X percentage, reducing or raising them to match their location on the curve of players abilities (ie if X is 20% and they have the lowest effective level then their skills would all be raised where necessary to be no more than 20% below the norm, if they were in the middle of those below the average then their skills would be raised where necessary to be no more than 10% below the norm etc). This would still make them not as good as the people they are playing against usually, but it won't make them so awful that its a boring embarrassment not even worthy of a Good Game post.

I'd put this at the top of the game issues needing addressing in GLB myself.

Originally posted by yello1
Originally posted by sam_acw

+ The AI needs to follow the inputs more exactly - there are too many occasions when the plays it calls are at the outsides of probablility. In particular the offensive AI calls more passes than you'd expect.

I am more concerned with the AI calling plays that are Flow Chart impossible to reach with the settings put in on the OAI or DAI.

Has happened more than once, verified by Mods.

Its frustrating to have that much time dedicated to making a game plan and then have it go balls up and cost you a TD or the game or the satisfaction that you at least DID something with all that time.



I've noticed that happening when I have a specific play included in my packages, even if I've taken out of the playbook. Packages seem to override playbook with that regard.

RE: QB's... My suggestion on the test server was the following....

Add in a final check. So, the QB will perform reads like normal, either the default for the play or the reads as determined in the custom AI. When he makes the throw decision, instead of actually doing the throw, a FINAL throw decision is made, based on a cone of vision centered on the receiver the QB decided upon.

The QB checks all eligible receivers in this cone to see if any would actually bee a better throwing decision, and if one is a better choice, he will throw to this receiver instead.

As it is, often the QB will throw to a receiver missing an obviously better choice in the same line of sight because of the way the progression works. EG, once he passes a receiver in the progression its rare he will go back to that receiver even though they may be very wide open when the throw decision is made.

I suggested that there be a vision check' to allow this last second receiver change, so its not automatic. Since its in a cone of vision, it doesn't cover every receiver on the field and thus while watching replays the 'laser line' will still be useful as you will have an idea of what part of the field the QB was looking at when he made his throw decision.

Cone could be larger or smaller based on the QB's Vision attribute.

Alternatively, each time the QB makes a 'progression' read, instead of reading the specific receiver, he should be reading all eligible targets in a cone centered on that receiver. In this case, the cone would likely have to be smaller, but the QB would be looking then at a section of the field, not just a specific target.

As it is, QB progressions are too robotic. I think that part of a good QB in Football is their ability to see more of the field at any one time. That in GLB there is a laser focus on a single player, then a laser focus on another player, ignoring the player prior, and so on like the normal progressions really does result in bad choices.

Along with this suggestion (since it would likely boost the passing game) I suggested that some of the various DB hampering be rolled back so DB's would behave more appropriately as well.