A lot of good examples there. I'll try to address them as best I can.
For which CB is on which WR, it currently does not try to line up slots. Much as there isn't a right/left designation, I didn't want to add in some functionality that you have no control over. Of course, all of this can be considered and improved upon, but for the base of 3.0, I decided it would probably be most fair to split chance of coverage amongst all the eligible defenders. When looking at all the possible ways of handling coverage assignments, this seemed like the most fair way to handle it short of setting which slot covers which slot from the defensive settings end which would add a whole new level of complexity to the settings that we probably do not want. So for now, if there are 4 WRs and 4 CBs, each CBs cover influence would be split amongst those WRs. Definitely room for improvement in this area, but this seems like the best way to handle it to start out.
I keep talking about influence, so it might help to understand what I mean by that. For receivers, they will be assigned a certain amount of influence for each location (VS, SH, M, L, D) that adds up to 1.0 (or 100%). This is based on the slot pass distributions, so if a WR was set to 10/20/40/40/50, his influence would be set to .06/.12/.25/.25/.32. Note that this is different than his chance to be the target in each zone. Basically these influence numbers represent the receivers chance to be in any given zone on any given play. Of course, a player can't be in EVERY location at any given time, but this would be similar to what pattern you would see over several plays of the same type. To determine the defenders influences, you would add up all of the WRs influence in a particular location and then spread it amongst the CBs, for instance. LBs get their influences from RB/TE influences. Safeties get influences set directly from the cover setting. What this all gives us is a way to pick a defender(s) in a certain location after determining the target location and target receiver. Once we have the target receiver and defender(s) we can determine how open the target is. This also means we can tell how congested a location is based on how much total influence is in the location. For instance, if all WRs are running deep, it is more likely there are more defenders deep as well. If there is more defensive influence in the location, typically due to S cover or having more defenders than targets, it will give the defense an advantage in cover. So influence is an important part of the cover logic.
Another important part to know - as you show in several examples - is that this is just the player's influence for COVER. Once the ball is thrown and the receiver is running, defenders will move towards the ball carrier. The chance they get to the ball carrier in a certain location is dependent on their initial cover influence but will change as the ball moves. So a S that was set to Long cover currently has some influence in the medium location and if the ball is caught VS (behind the line), he will probably be able to react to try to take part in the tackle attempt by the time the ball carrier gets into the short location. Now if the ball was caught in the short area instead of the very short area, there's actually less time for him to react to be part of the play, so he most likely would not be tackling a receiver for 4 yards if the receiver caught the ball in that area.
The nice part is that most of this is isolated, so we can adjust those initial cover influences and also adjust how well a player would react to take part in the tackle after the catch. Most of this is kept pretty simple right now with ratings mixed in where I can, but this is another area that could be improved. I could definitely see GI play a greater part in the S moving up from Long cover to tackle a guy in the short location than what it is now, but I'd like to get the base pretty solid before introducing more complexity. It's already difficult enough to balance all the moving parts as it is. But definitely things we can improve on later.
Oh, as far as what happens when there are 4 WRs deep and 1 S, I don't think I addressed that. Basically when looking at the cover the defense has in a location, we look at those influence numbers and can look at the single value split across receivers. So the defensive influence advantage might be 1.5 with 2 WRs and 1.25 with 4 WRs in the case with 1 S in the area. What this tells the engine is that there is less chance the S is covering the target when there are 4 WRs in the location. Again, lots of possibilities to work player talent into this.