I guess the way to look at it is that we should not see another anytime soon.
The bad thing with such "bad luck" events is that past results don't affect future ones.
With the huge volume of blocks your pool finds can you give a probabilty of invalid blocks occuring in close succession based on your past results?
Among other statistical things, it's also dependent on the 'bitcoind' connection to the rest of the bitcoin network, the number of connections it has and also issues like the network delay from 'bitcoind' to those it's connected to.
It also doesn't help if the miner that generated the block is slow to notify the pool or the pool has ANY delay in notifying the 'bitcoind'
So basically, each pool will have different expectations.
Just in reply to the above post:
My "
Among other statistical things" means of course the probability of 2 miners generating a block about the same time which is of course totally independent of the pool.
And of course to clarify something that anyone might be thinking (though hopefully not) there is no way for Inaba to take advantage of making an invalid block.
Either the block is valid and everyone (including Inaba) profits from it, or the block is invalid and no one profits from it until the next block.
The person who generated the Invalid can easily check if it was a real block so there would be no reason for any pool to pretend it was an Invalid
Aside: well easily if they use cgminer, I'm not sure how other miner programs notify you of block, but cgminer does with the block hash (I wrote that code

)
Edit: yes of course this does fall directly into the area of bitcoin that is 'random probability'