If you're clever about how you blend models you can pretty much ensure that the performance of the (weighted) averaged model is strictly better than the (weight

If you're clever about how you blend models you can pretty much ensure that the performance of the (weighted) averaged model is strictly better than the (weighted) average performance of the models. And increasing the space of possible models pretty much guarantees the performance will improve, provided you can still find a good enough optimum.