Whether or not they would be more effective if they shifted focus to be more defensively responsible (even assuming that such a transition could be made) would depend on the rate at which they're trading goals for for goals against. In order for such a shift to be an improvement, goals against would need to go down by more than goals for would.
On one hand, we might think that there are diminishing marginal returns for cheating on offense—going from cheating 60% of the time to 70% of the time gives a bigger increase in goals for than going from 70% to 80%. As such, we might think that shifting away from this degree of cheating could lead to a greater decrease in goals against since going from defending 20% to 30% would be more beneficial than going from defending 30% to 40% (if we also believe there are diminishing marginal returns in defending).
On the other hand, this tradeoff depends on the relative size of the effects. And if that line is bad defensively, the efficiency at which they are converting defensive effort to fewer goals against might be low. Lower even than the smaller increase in goals for you might expect from cheating even when you're already cheating a lot.
But that's probably just a long winded way of saying what everyone already knew.
"If you torture the data long enough, it will confess to anything." - Ronald Coase
"[...]all models are wrong, some are useful." - George E. P. Box