EA Conference Call With Sean Bailey, Lead Designer - Gamplay

[CowherPower]

Quote: Originally Posted by jwb0581

I coach football for a living and this is how it works. You watch film of your upcoming opponent and chart what they do out of each formation. Say you're playing a team that ran I-twins right on 25 occasions on the film you have. 23 of those 25 they ran a power running play to the left. The other 2 times they ran a power running play to the right. You do this for every formation. You then gameplan and implement the plan during the week at practice. If they come out in I-twins right we will run a 46 look with our safety hammering down. We will mix it up slightly but anything we call will be a run-oriented defense, with some wiggle room for down-and-distance. We will tell our defensive players to watch run to the left out of this formation. We will watch what they do in the game. At halftime we will adjust what we're doing vs. I-twins right if the 1st half has indicated that we need to. As an offense, sometimes we will purposely show the same play over and over out of a certain formation, setting up the defense for a play when we need it. We will show a tendency and then break it in a high-leverage situation. If the CPU in NCAA could tally what plays are run out of which formations during the season, and adjust even on some basic level, the gameplay would be revolutionary.

Amen,great post.

[Bootzilla]

Quote: Originally Posted by CowherPower

I would appreciate it if you didn't try to put words into my mouth. I NEVER,I repeat NEVER said anything about the cpu defense calling different plays. The cpu will call different plays,but will NEVER adjust to what I'm doing....period! If the a.i never adjust's to one's playcalling tendencies,how will the outcome "vary"? I would STILL be able to run the same play 10X in a row with success,no matter what the defense chooses to pick.....4-3,3-4,nickle,dime or whatever.

Dang dude you really got upset about what I said. Wow. I don't understand how you can say that the outcome of a play will not vary depending on the defense that is called. Heck, the defense can luck up into a right call. Right? If you run a play 10x's in a row vs. 10 different defenses that would suggest to me 10 "POTENTIAL" different outcomes. If you go to practice mode and run the same play against different defenses you'll get different results. Or at least I do. I'm not putting words into your mouth. You say you WILL be able to run the same play 10x's in a row with success. I say how do you know?

[Bootzilla]

Quote: Originally Posted by jwb0581

I coach football for a living and this is how it works. You watch film of your upcoming opponent and chart what they do out of each formation. Say you're playing a team that ran I-twins right on 25 occasions on the film you have. 23 of those 25 they ran a power running play to the left. The other 2 times they ran a power running play to the right. You do this for every formation. You then gameplan and implement the plan during the week at practice. If they come out in I-twins right we will run a 46 look with our safety hammering down. We will mix it up slightly but anything we call will be a run-oriented defense, with some wiggle room for down-and-distance. We will tell our defensive players to watch run to the left out of this formation. We will watch what they do in the game. At halftime we will adjust what we're doing vs. I-twins right if the 1st half has indicated that we need to. As an offense, sometimes we will purposely show the same play over and over out of a certain formation, setting up the defense for a play when we need it. We will show a tendency and then break it in a high-leverage situation. If the CPU in NCAA could tally what plays are run out of which formations during the season, and adjust even on some basic level, the gameplay would be revolutionary.

I get how it works in real life. I'm giving away my age but, I played D-1 and both of my kids are current D-1 athletes. I coached DB's on the high school level for 10 years before switching to track & field and basketball. I currently have 6 kids that I've coached in AAU at D-1 universities. I know all about making adjustments. The only points I was trying to make is that in "virtual football" down and distance tendencies is not a game killer because results can/will vary based on what the defense calls. That's all.

[Village Idiot]

I'm with ya, Boot.

It's not like you ever said the game shouldn't have adaptive AI, just that the fact that it doesn't shouldn't equate to being a game-killer. I completely agree.

Although I think this thread, and the other AI thread, shows that adaptive AI should in fact be the #1 priority for 2010.

[jwb0581]

Quote: Originally Posted by Bootzilla

I get how it works in real life. I'm giving away my age but, I played D-1 and both of my kids are current D-1 athletes. I coached DB's on the high school level for 10 years before switching to track & field and basketball. I currently have 6 kids that I've coached in AAU at D-1 universities. I know all about making adjustments. The only points I was trying to make is that in "virtual football" down and distance tendencies is not a game killer because results can/will vary based on what the defense calls. That's all.

An impressive resume, to be sure.

Down and distance tendencies are ok. What's not ok is being able to run, for example, I-formation, 2-tight end play action to the lone WR running a streak all day long without the CPU adjusting. I know there is a slight chance they will luck into the right defense to stop the play about 1/10 times. But if it works 9/10, or even 5/10, it's totally unrealistic because there are no adjustments.

I know that you know what I'm saying. I'm just trying to make it more clear for everyone.

[Bootzilla]

Quote: Originally Posted by jwb0581

An impressive resume, to be sure. Down and distance tendencies are ok. What's not ok is being able to run, for example, I-formation, 2-tight end play action to the lone WR running a streak all day long without the CPU adjusting. I know there is a slight chance they will luck into the right defense to stop the play about 1/10 times. But if it works 9/10, or even 5/10, it's totally unrealistic because there are no adjustments. I know that you know what I'm saying. I'm just trying to make it more clear for everyone.

I totally agree. Just putting a positive spin on something that I wasn't all that happy about. Since I mix up my play calling anyway, it would have been nice to have never even heard this. I would have mixed up my calls and remained blissfully ignorant.

[Bootzilla]

Quote: Originally Posted by Undefeated

I'm with ya, Boot. It's not like you ever said the game shouldn't have adaptive AI, just that the fact that it doesn't shouldn't equate to being a game-killer. I completely agree. Although I think this thread, and the other AI thread, shows that adaptive AI should in fact be the #1 priority for 2010.

You're on the same page as me. Could not agree more in regards to 2010.

[countryboy]

Quote: Originally Posted by jwb0581

Down and distance tendencies are ok. What's not ok is being able to run, for example, I-formation, 2-tight end play action to the lone WR running a streak all day long without the CPU adjusting. I know there is a slight chance they will luck into the right defense to stop the play about 1/10 times. But if it works 9/10, or even 5/10, it's totally unrealistic because there are no adjustments.

very well put.

[callmetaternuts]

Im a little ticked that adaptive AI (to any degree) isnt in, but alot of these things havent changed fromt he last few years. We did without them in the past, we can do without them now, but lets make our push as a community to get them in next year.

[Dbrentonbuck]

Quote: Originally Posted by jwb0581

I coach football for a living and this is how it works. You watch film of your upcoming opponent and chart what they do out of each formation. Say you're playing a team that ran I-twins right on 25 occasions on the film you have. 23 of those 25 they ran a power running play to the left. The other 2 times they ran a power running play to the right. You do this for every formation. You then gameplan and implement the plan during the week at practice. If they come out in I-twins right we will run a 46 look with our safety hammering down. We will mix it up slightly but anything we call will be a run-oriented defense, with some wiggle room for down-and-distance. We will tell our defensive players to watch run to the left out of this formation. We will watch what they do in the game. At halftime we will adjust what we're doing vs. I-twins right if the 1st half has indicated that we need to. As an offense, sometimes we will purposely show the same play over and over out of a certain formation, setting up the defense for a play when we need it. We will show a tendency and then break it in a high-leverage situation. If the CPU in NCAA could tally what plays are run out of which formations during the season, and adjust even on some basic level, the gameplay would be revolutionary.

I don't see how this would be that difficult. It would be tied to your profile. The game already keeps track of a lot of stuff you do in order to give you user stats. Why can't the AI then benefit from that tracking and know that I like to run up the middle more than to the outside and that I like to pass on crossing and post routes more than streaks and outs. The AI should know that I probably have a close to 55/45 Pass to run ratio and so should play a balanced defense concentrating on the middle of the field for both run and pass. Didn't 2k do something like this when they implemented the player profile thing where it would learn how people played and you could then play against the AI version of their style?