In attempting an intentional, non-obvious walk it is very likely that at least one of the pitches will be hittable, thus defeating the purpose.
Define 'hittable.'
Seriously. Will there be a new 'almost strike' zone that is not a strike, but an 'unintentional' ball?
Geez, this opens up the door to not only bean balls, but catcher interference, and maybe even intentional balks. All so that a team sport can be made even more the domain of individual athletes.
But that's easy to fix! You just have to make a bean ball an automatic two-base walk! Mission creep!
Look: I said how to do it. I gave no opinion on whether it should be done...
:-)
Others have doubtless pointed out that the DeLong solution has the problem of disproportionately penalizing naturally wild pitchers. I'm more interested in whether the Bonds intentional walk phenomenon really *is* rational management. In the Historical Baseball Abstract Bill James shows fairly conclusively that *always* walking a player is a bad strategy, no matter how good that player is.
(Brief methodological note: James amped up Babe Ruth's best season, then put the augmented Babe on a team of complete stiffs, and ran paired simulations of 1,000 seasons. In the first group, the babe hit away, in the second, the computer manager gives him the free pass. Verdict: it doesn't work. The IBB babe teams score more runs, with the hitter behind Babe averaging something like .250 with 9 HR and 152 RBI.)
I'd love to know how James' analysis would have played out with a few more rational management strategies -- always walk Babe with a man on second; always walk the babe with two men on and two out, etc. I suspect we'd find that currently managers are walking Bonds too much.
Surely some sabremetrician has done this analysis. I'll go look.