Your feedback elements are all good, except I don't understand the role of the weakened ocean/and contrast (could be a gap in my understanding). I do not usually think of the convection at the dateline as the first step, although it could happen that way. I think "all together" is the best way to conceptualize the feedback loop, although it is not entirely simultaneous in practice and is instead incremental and staggered. I usually think of the wind anomalies and the SST changes as the two most fundamental components, but all of them you mentioned are at play. I think of the convection as a result of the increased SST, so think of it as at least 3rd in the order, but there is not set rule about that. Increased convection would change the atmospheric circulation, which would change the SST in that same direction but more so, etc. About 2014, one factor that stands out in my mind is that the westerly wind bursts and the amazing increase in OHC occurred a little too early (in March or so). To a moderate extent, the ENSO phenomenon is seasonally phase-locked. It must get revved up near the time of the spring "barrier", April to June. May is the center of the barrier. If what had happened in early spring 2014 happened a bit later, when the SST is at its climatological peak and anomalous convection is easier to get, we may have had a big El Nino in 2014-15. But most of the critical fireworks happened a couple of months too early when the baseline SST was still a bit too cold.