The QB read progression absolutely must be fixed. The I'm going to stare down a WR while he finishes running a route even though three other WRs are wide open is the stupidest thing I've seen in a football simulation.

Example 1: http://mfn19.myfootballnow.com/watch/1475#281914

Look at the WR at the bottom of the screen. QB will never see him because he's waiting for another WR to finish his route. This is not a FOV problem either. The other WR is just running a longer route so the QB waits patiently.

Example 2: http://mfn19.myfootballnow.com/watch/1475#281957

The slot WR makes a beautiful break to the outside. Which the QB would see if he wasn't staring at the WR running the deep route. Instead the QB patiently moves outside the pocket and waits to be sacked so the WR can finish his route. Again, not an FOV problem. The QB will never use the FOV skill because he has to stare down every WR as they finish the route.

My favorite play of this game was when the QB threw the ball away on 4-10 while being down by 7: http://mfn19.myfootballnow.com/watch/1475#282035

I don't mind losing close games. But I mind losing when the QB read progression code is so horribly broken that you might as well only have one WR running a route on a play.

The QB (based off of his attributes) should be able to know what routes his guys are running, read if the defense is in man or zone, read the safeties and determine what his best option is. For example the defense is playing a cover 2 shell and you have someone running a deep post or a seam route, those guys should be good options to throw to. You wouldn't want to throw that fade, streak, deep out or flag route to the outside though. Or if the defense is playing a cover 1 then you would probably want to look at something like a corner route. The problem it seems to me with the passing game in this sim is that it does not really seem to read the defense. It just simply waits for the primary WR to finish his route and then decide where to go with the ball.

This is exact what happens.

http://mfn19.myfootballnow.com/community/1/1321?page=2&x=1buknm&y=bvw#7802

I dunno... Your QB is a rookie and has an intelligence of 15, and he made a couple of bad decisions... heh!

At a certain point, you have to realize that this is a simulation. You're upset that your qb threw the ball away on 4th down, yeah maybe not a great thing to do, but he's not actually an actual human with an actual brain. There are, and always will be, limitations to AI. It seems to me that you are expecting your qb (again, rookie with 15 int) to make Peyton Manning like reads and decisions. Perhaps you are asking for too much from the game.

I'm not mad that my QB threw the ball away on 4th down as much as I am that he took three sacks by staring at a WR.

You're also overvaluing intelligence by quite a bit. If Intelligence actually affects a QB's read progression, it's not documented. The only thing that the documentation says about intelligence is this:



I'm actually well aware that this is a sim. As such, I'm well aware of how game code breaks down when certain variables are introduced that outweigh all the others.

For instance, here's how the QB read progression code works:
1) QB randomly chooses a receiver to be the primary target based on the weight of the play and the routes being run
2) The QB waits for the primary receiver to complete a route and then moves on to the next route, waits for it to be complete, etc.
3) Based on the receiver that the QB is staring at, he decides if he wants to throw the ball

Intelligence, therefore, has very little impact on QBs. The only skills that really affect the pass are accuracy, arm, and to a lesser degree lookoff. Given how bad the OL code works and the severe penalties to throwing under pressure, scramble and acceleration are next.

In the old RPG language, if you min-max those five skills, it doesn't really matter if your QB has a zero in intelligence and FOV. That is how you break code.

On here I get to do it for fun. Most of the time I get paid to do it.

How the QB determines the primary WR is here: http://mfn19.myfootballnow.com/community/3/1361?page=0#8134

Intelligence does influence each of these decisions. FOV also influences his decision to throw the ball if there is a more open receiver in his FOV that matches the objective of the play.

You are right, though, in that it is very simplified, as the game matures everything will become more complex, but my goal is that those complexities must come with the ability for you to control them to some degree and also have feedback on why the player didn't do what you asked them to. There must also be the ability to counter these controls. For example, an earlier version gave the read progression more in line with who was the most open receiver. Because defenses in general don't have a great way to cover the tight end routes, all of the leagues #1 receivers were tight ends. Giving you the ability to target your tight end as the primary receiver must also come with a way for defenses to counter.

I'm still working to get the baseline decision making solid before extending the game planning into these controls; right now, the simplicity is this: if you call a long pass, the QB will most likely target a long route. If you call a short pass, the QB will most likely target a short route. The high sack count can be dealt with by giving the QB better baseline decision making.

FWIW, in general, pretty much every decision in the game is passed to some degree through the intelligence and experience filters. The 'handful' comment is more directed at decisions where intelligence plays a key role in the decision.

I agree that at this stage in the game that simplicity is best. Additional complexity of the algorithms will add additional performance concerns when multiplied across all of the leagues.

My goal of the quoted post is to figure out how to do the best with the code right now.

My goal of the OP was to highlight some of the places where places in the code detrimentally affect other places in the code.

I'm still thoroughly enjoying playing the game. (And part of my enjoyment is because I like seeing what happens to software when you throw all sorts of random variables at it.)