That is your argument sure, and I agree with you, mostly. The problem is kyear thinks ATG games will be lower scoring. If that's the case, even if the statistical gap between Imran and Marshall reduces, there's arguably just as much value added by Imran's batting in ATG scenario as in real test matches.
Imran averages 38 and Marshall 19 in real tests, in which, let's assume 300 is an average score. If we assume in ATG games that 200 is an average score, Imran averages 27 and Marshall 15 or something, those 12 extra runs are now extremely valuable because of the assumption that the games are lower scoring and even the ATG batsmen aren't delivering their usual numbers.
I just don't think "ATG XI games will be low scoring" and "ATG XI batting depth doesn't matter" are positions that you can hold together. There's a strong argument that the marginal loss in bowling quality in games that are expected to be low scoring is not as important.
The reason I believe these games would be low scoring is precedence.
In cricket, batsmen stats are more impacted by great bowlers, than great bowlers are impacted by batsmen. I should qualify that by saying great fast bowlers. Look at Lara, Tendulkar, Richards, Sobers, literally all of them.
And for me the batting depth matters less for two reasons.
1. I would go for the absolute best bowlers with the possible exception of the no 8 position, and that person, would be Hadlee over McGrath, but then I also believe McGrath is the 2nd best pacer ever and him and Marshall would bowl well together, and I also would prefer to have guys from different eras. Hence Marshall, Steyn, McGrath is perfect for me.
2. Diminished or depreciated value of the secondary skills. And let's be honest, if Bradman, Sobers and co have been skittled out, don't see any tail making much of a comeback.
I want the best bowlers to be able to bowl them out for as cheaply as possible.