Few Aussies seemed to worry when McGrath, Gillespie and Warne were backed up by Lee and no other. I'd guess that almost all the most successful WI attacks of the past were also 4-man attacks (even when Walsh and Ambrose had to carry pretty ordinary support bowlers).

The argument based on wickets per match is also completely spurious. Aside from anything else, wickets per match is dependent on overs bowled, who they are bowled to and the like... but more importantly, you have to take someone off in order to put another bowler on... so adding a bowler with strike rate x to a side that already has bowlers with strike rates a, b, c and d is rarely going to make a major difference.

These bowlers cannot bowl simultaneously, and few sides can manage even 90 overs in a day: what matters is that in those periods of the day where bowlers are dominating, a high-ish strike rate option exists... and in those periods of the day where the batsmen would be favourites with almost ANYONE bowling, the bowlers can build pressure and take control in ways that tend to be reflected in averages and economy rates.

A side will probably need to take 20 wickets in at least one match of a series to emerge victorious (though declarations might lead to success with fewer wickets). In most games (and most conditions), no bowler (or collection of bowlers) is going to run through any decent top order: what matters is simply capitalising when an opportunity arises - with the new ball, when overcast conditions allow swing, when the ball is reversing or when cracks / rough allow exploitation of the pitch.

Most matches against GOOD opposition should be drawn: for either side to lose should require a pretty abject performance with both bat (giving wickets away) AND ball (leaking runs). What matters is being able to capitalise in those few matches where a result can be forced.

The attack needs to be able to both ensure the bowling doesn't contribute to a loss in the games that cannot be won and also shift the odds in the direction of a win where the pitch / conditions / opposition present an opportunity. For that, you need 3 good (dependable) bowlers and someone else who can offer something a bit different - but for the reasons I gave earlier, adding a further option is not really shifting the odds in most situations.

I also find the workload arguments spurious: bowlers are only overbowled when a bowler breaks down and the occasional bowling sucks... and that's rare. For England it's just not going to happen: the occasional bowling could cope with a far greater workload in the unlikely eventuality of a leading bowler being removed from the options list.