OK, let's continue with the discussion ..
Why should I operate according to MEL for gear U/S, just because I delay retraction?
What does your AFM say in respect of selection of gear during the takeoff ? Commonly something along the lines of "retract gear three seconds after liftoff .. ?"
If the pilot wishes to operate in a manner significantly different to that specified in the AFM, which may/will compromise the certification basis, then one needs to consider what answers the pilot might have to questions at the Enquiry after the mishap .. regardless of whether the point in question has any relevance to the mishap.
No different to, say, using a non-AFM flap setting for takeoff .. eg nil flap on light twins .. which I routinely did in the past. For the twins I routinely flew, I had done sufficient climb testing to provide me with the basis for the variation and, more importantly, the basis for an argument in court.
In essence .. in the absence of an abnormal/emergency, either you do
(a) sufficient certification work to establish compliance and a defence at law or
(b) follow the AFM or
(c) use the MEL as your get-out-of-jail-free card
Just to say that you will do what you want to do today because you are the pilot is a rather reckless approach to risk management, I suggest.
I have to delay flap retraction to 4000 feet ... Engine failure procedure will have me retracting them at 1900 feet (1500 feet AGL). Should I in this case check the MEL for restrictions for flying with flaps U/S?
A facile argument. The takeoff will (or should) have been predicated on
(a) the AEO requirements, AND
(b) the OEI escape as appropriate.
A pilot (generally not the reputable airlines) without benefit of a rigorous ops engineering department takeoff procedure in support who departs without an AEO to OEI transition plan to cover all eventualities eventually may not have a good day .... He/she certainly won't have an easy time of a subsequent Enquiry.
the performance manual is a tool to support the pilot in making good decisions.
No. The performance data is to provide the pilot with information to permit compliance with the rulebook. The pilot should be making sensible decisions within the parameters of those requirements for other than emergency situations beyond the normally expected planning regime.
Alternatively, one could pose the question .. how does the pilot, in the situation you posit, make good decisions whilst winging it on the fly ? I put to you that a good decision requires both a successful outcome (in respect of the decision process) and a defensible logic .. the latter including compliance with the Design and Operating Standards where operations, essentially, are normal.