Thanks. My question is really about CPU AI. We've heard some things from Josh Looman about CPU AI when it comes to roster management, but for Mike Young, I'd really like to know more about in game AI.
For in game CPU AI, here are some things I'd like to know:
- Will the CPU try to make first downs? In M13, the CPU has no awareness of the importance of moving the sticks, so even if you adjust sliders to favor the CPU more to help it sustain drives, you still have to try to over power it a bit to get more intelligent play calling and play execution. Also, CPU receivers often run out of bounds before the 1st down line.
- Will the CPU be able to give any weight to the weak links on the user controlled team to maybe target a weak DB, for example, or make the receiver to that side the first read?
- Will the 1:01 time out bug be gone?
- Can the CPU have any ability to audible out of a play if the defense is, say, looking like it's selling out tot he run?
- Can the CPU adjust in game to favor, say, passing, if early success in the passing attack looks better and the run is getting stuffed (or vice versa)? Or will the CPU stick to a rigid game plan no matter how the game is going?
- Of course, will the bad run/pass balances be fixed across teams, especially for teams like the Packers and Patriots?
- In CFM, if the Jaguars sign a highly rated Andrew Luck 4 years in, will they still run a play book that mostly runs and dumps off short passes, or will the CPU be able to adjust pass/run balance, say, in the off season, based on that season's personnel? In other words, can the CPU recognize that Andrew Luck in not Blaine Gabbert and adjust its playbook or play calling accordingly?
- Will the CPU do a better job of emphasizing the passing attack when behind before the last 2:00 of the 4th quarter? In M13, if the CPU is down, say, 2 touchdowns with 8:00 remaining it will still default to its regular play book, running frequently, and passing short and medium. The CPU should be taking more chances when behind. It does this in the 2:00 drill, but by then the game may be completely gone.
I know you can't probably get to all of those. A common element of all of them is whether or not the CPU can do any dynamic assessment of play calling based on in game situation (down and distance, or time and score). There's also a little bit of playbook adjustment to personnel, at a minimum, the CPU's own personnel. You get the idea.
Thanks!






Comment