Hello Stephane,
what exactly would you like to accomplish with this solution? Better executions/lower slippage?
The problem of this solution is that before the first order is entirely filled, the price of the second leg can already drift away and you may enter at worse z-score.
Also the implementation is a bit tricky, for instance - how long should the app wait for the first leg to be filled? What to do, if it takes too long?
But this is good question, let's open a discussion here.
(Btw for enterprise clients we have a partner providing smart execution algos for now - the Algotrader platform.)