The number one problem in WITP A2A model is that every fighter unit will attack every enemy fighter unit, then every remaining intercepting unit will attack every enemy bomber unit. Not the way it was working in real life....
Also we saw huge raids mixing Swordfishes to P-38 flying together at long range with no problems. They should arrive mostly in smaller raids.
Same thing for raids with hundred of bombers of the same type (ie B-17) concentration should take time, so for each 50 bombers range should be reduced by one. Also there should be a check to see if the concentration fails or not.
As for the CV, every CV raid/battle saw the CVs of both sides launch several waves. The reason is simple, a CV carried more AC that she could launch in one group. This concept of "deck load" is ignored by the game. It seems to me that usually a CV was able to launch about 40-50% of its load in one raid (the other AC being used for CAP and patrol)
Of course every of the points above is reducing the size of raid, but the CAP efficiency should be greatly reduced. I read an interesting figure yesterday. On D-Day, the RAF fighters units covering the invasion force flew 1567 sorties (only couting LRCAP over the fleet and beaches). That was to only have 72 AC on CAP at any time (ie 5% of the total of sorties), ie 36 over British beaches and 36 over US ones. They were able to intercept the dusk raid by the Ju88 of I./ZG 1 but the well-known strafing pass by two pilots of JG 26 or far-less-known rocket and strafing attacks by JG 2/26 later in the afternoon were not intercepted.
It seems to me that of the fighters allowed to CAP, a maximum 30% should be on CAP above their base/ship at any moment, 30 other % may have a chance to scramble if the raid is reported incoming and the other 40% will be refueling, rearming, under repair, returning home with few fuel, etc...
Again the 30% above is consistent with every account I have read of Pacific battles, where CAP over US or Japanese CV was listed as being 6-8 fighters per CV rather than 25-30...
Then for the A2A model, here is an algorithm for an escorted raid:
1) escorting fighters will be divided in groups of 4 or less fighters
2) depending of their agressivity, number relative to the escort, speed difference vs their escort, each of these group will be considered as close escort or "sweep"
3) the defending fighters are also divided in groups of 4 or less fighters
4) depending of the radar warning, speed of the incoming raid, state of the ship/runway, a number of fighters may be scrambled. They are also divied into groups of 4 or less fighters.
So we have "sweep" and "escort" fighter groups on the attacking side, "CAP" and "scramble" groups on the defending side
The first phase of the battle saw each "sweep" trying to engage enemy fighters.
1) first they have a chance to attack a scramble group. If they manage to engage one, a CAP group will try to engage them and so cover the scramble group. In this case the battle will then be fought between the CAP and sweep group on even terms, if not the "scramble" group will be attacked with a strong disadvantage.
2) if a "sweep" group hadn't been engaged yet, it will try to engage a CAP group
The next phase, see the "CAP" and "scramble" groups that have not yet been engaged trying to engage enemy bombers
1) each unengaged "CAP" group will try to engage an enemy bomber unit. If they manage to engage one, an escort group will try to engage them and so protect the bombers. If they succeed, the battle is fought between fighters, if not the defending fighters have a go against the bombers.
2) then each unengaged "CAP" group will try to engage enemy escorts.
3) then each unengaged "scramble" group will try to engage the bombers on the same way, but with less success chances and more chance for the escort to stop them.
4) then each unengaged "scramble" group will try to engage enemy escorts.
5) each unengaged escort group may try to engage an enemy fighter group, depending of their agressivity, in this case they will become "sweep" groups.
At the end of the phase, scramble groups become CAP groups. Each damaged AC may become a group alone, or be covered by a wingman. Each engaged group will have a chance to have no more ammunitions, fuel and will try to leave the battle area. If it didn't succeed (depending of their speed relative to enemy speed, damage and cloud cover), it will be a easier target the next turn.
If the bombers are especially slow, and depending of the radar warning and cloud cover, this phase may be repeated, allowing more attacks and more defending fighters to be scrambled (but the overall total should only by 30% or scrambles)
Then there is the bombing attack, as it is now.
And then I would like a pursuit phase. Each remaining "sweep" fighter will become an escort and each defending fighter will have a chance to attack.
For a sweep mission, there will only be the sweep phase of the fight, with a phase on the approach, a phase replacing the bombing and a pursuit phase.
It may seem complicated, but it is an easy algorithm to put in place. Of course, the success chance and formula should be worked on to achieve historical results... Each aircraft may have a chance to engage 1-3 enemy formations (or 1-3 times the same), with something as 25-50% each turn to engage one. Then depending of the experience of both attacker and target, performance of Ac and previous damage, a given pilot may fire at and down several AC in the same phase.
I like Admiral's ideas/concepts.
I would like to suggest that rather than the 4 plane groupings that the Admiral specifies that some accounting be done for the air to air fighter doctrines of the different countries.
It seems to me that the US Navy (anyway) had adopted the Leader/Wingman concept and tried as best as circumstances allowed to use that as its base tactical unit in air to air combat. I believe the Brits early on in the war used a 3 plane vee as their base tactical unit but at some point subsequent to the BoB adopted the same idea as the USN (and their German opponents who seemed to been first off the mark with the idea). Whether or not this practice had migrated to their units in the Far East by Dec 41 is unknown to me. The three plane vee was harder to control and easier to break up though which of course explains why it fell out of favor.
In a large variety of readings it seems to me that the IJN did not have a formal standard tactical unit that was designed to function in combat. The individual pilot seems to have operated as a "lone wolf" throughout a significant portion of the war. That's not to say that they didn't help one another in combat but that there was not any formal setup where one pilot was responsible for any specific other pilot as is established in the Leader/Wingman concept.
Evidence of this can be found in "Shattered Sword" where the operations of the IJN CAP over KB are very exhaustively described. Descriptions such as the following are made "PO1c xxxx led 3 fighters of the Akagi's "4th watch" into the air, the 2 planes of Ens yyy's Shotai attacked the TBDs, PO2c zzz landed to rearm while the other 2 planes in Hiryu's 2nd watch continued their patrol...Kaga launched the 5 Zeros of its 3rd watch".
Shattered Sword uses the word "organic" to describe tthe IJN CAP and seems to indicate that it reacted more instinctively than thoughtfully to the USN attacks at Midway. Though it brought overwhelming firepower to bear on the individually attacking squadrons from Midway and Torpedos 8 and 6 this same instinctive reaction led to a serious problem when confronted with the high/low and escorted attack of Yorktown's stike: far too many CAP pilots got into a furball with the escort than was required to neutralize it and ALL those pilots who didn't went after the TBDs leaving Yorktown's SBDs to make their attack utterly unmolested. Further evidence of the instinctive and "lone wolf" philosophy of the IJN fighter pilots is indicated by the attack on some straggling SBDs by most or all of the escorting Zeros for Hiryu's first strike on Yorktown. In the greater scheme of things it is hard to justify them abandoning their charges (the bombers), even temporarily, when any hope of victory rides on their ability to strike the enemy. As it was the attack on the SBDs, deprived the escort of 1 or 2 Zeros and cost the Vals several bombers shot down before the remaining escorting Zeros were able to engage the CAP. As I recall a similar incident occurred at Santa Cruz and though in that instance it happened before the USN planes had expended their ordinance, it also deprived the IJN bombers of escorts before they encountered the US CAP.
In any case, it would seem that rather than a 4 plane grouping IJN CAP and escorts, up to some point (randomly determined maybe - at some point I believe the IJN officially started to operate in leader/wingman mode) might be better simulated by a randomly generated number of say 1 to 5 planes per group while USN would be in multiples of 2 (4 max) while the Brits and maybe Dutch would be in 3s.
Well just some thoughts.