Interesting thoughts in this thread .. my 2c worth .. much of which repeats/summarises bits and pieces from various posts ..
(a) consider going for the appropriate OEI pitch attitude on liftoff for a light twin .. gives a nice AEO acceleration and, if we have an engine failure, control it, do the necessaries and then finesse the resulting airspeed a bit. On higher performance machines, there may be a need to pitch higher to cover the AEO case .. with a need to lower the body attitude during the initial OEI controlling, although this is more relevant to FAR 25 twins.
(b) at low weight/pressure height/OAT, any aircraft should go OK on one. As we approach the other end of the envelope, the available range of speeds to achieve a climb on one reduces markedly. The climb curve is bell-shaped .. if we are above the peak of the bell, then it is downhill all the way until we have a bit of excess grunt left over. For most light twins, at gross weight and a bit above sea level, we need to be somewhere near blue line (as corrected for the conditions of the day) and in the appropriate configuration to get much of the little performance which may be available. Too slow or too fast and we go down rather than up.
First priority for a continued takeoff ought to be stay right way up, then climb or maintain height as possible or appropriate, and then make sure the aircraft misses the rocky bits ... eventually getting back to a suitable runway.
(c) best performance bank OEI is around 2-3 degrees to give zero sideslip .. doesn't matter what multi is being discussed, Concorde included .. wings level is easier to fly for a marginally degraded climb performance .. take your pick. 5 degrees is relevant to Vmca determination and initial OEI control if you are near Vmca .. as the speed picks up, the 5 degree thing is largely irrelevant .. and, if you are near Vmca in a lightie with a failure .. and near to the ground ... then the choice of options is very limited. Climb performance wings level and at 5 degrees generally is similar, so why bother with the additional workload of flying with a bank of 5 degrees for no gain ?
(d) the AFM figures and recommendations are based on the presumption of critical conditions .. ie minimum available runway length and a need to achieve whatever climb performance, if any, that the particular design standard looks for. If you have more benign conditions, then the available range of options increase. Apart from considerations of standardisation (which are very important, especially in a larger operation) ... it is a bit silly, from a risk mitigation viewpoint, to adopt procedures intended for a critical situation as being rigidly appropriate for ALL circumstances where other options may offer a lower risk scenario.
From a simple energy viewpoint, height is more useful than speed .. but energy considerations are only part of the flight management equation...
For instance ... taking off from a 12,000 ft runway with little in the way of terrain to consider ... lifting off at minimum AFM speed, having one quit, experiencing the "oh dear" time delay, slowing down, rolling inverted, experiencing the resulting crash-burn-die scenario is no more/less silly than trying to hold it on until blue line on a short runway with terrain considerations, having one quit just before blue line, unsuccessfully rejecting and falling into the rock-filled gully beyond runway head with a slightly different crash-burn-die scenario ... The best option needs to be considered on a day by day, runway by runway basis.
(e) there are no guarantees at all ... FAR 25 included .. as shown by a succession of crashes over the years ... certification data give us something in the way of reference information to work with .. the outcome on the day depends on a mix of training, knowledge, luck, skill, planning, and keeping cool under pressure.
One thing is certain .. get too slow and over you go .. so the Vmc departure spectre needs to remain very much in focus for a mishandled engine failure and continued takeoff.
The aim is "don't crash".
I haven't flown any light twins for quite some time but, in years gone by, if in a Shrike or Chieftain etc., with a comfortably long (and reasonably smooth) runway, zero flap takeoff with a rotation somewhere approaching blue line seemed a reasonable compromise to me and certainly simplified the continue/reject decision ... none of the old DCA examiners, whom I briefed appropriately as to TOD corrections, conservative assessments of accel-stop distances, etc... ever queried my flight management philosophy on renewals ... (Some aircraft, of course, weren't appropriate for a faster rotation speed.... the dear old Twin Commanche is the best example which springs to mind). Conversely, on a shorter runway, the P-chart Vtoss speed makes much more sense.
My point is that we are faced with death from a number of quarters during a takeoff ... the best solution seeks to stay as far away from all as best as one can orchestrate for that particular takeoff ... generally a one-size-fits-all approach is not the optimum philosophy.