The latest ENSO strength outlook continues to indicate a moderate probability (~35%) that the peak Nino-3.4 index will fall below -1.0C and a low probability (~10%) that it will fall below -1.5C. Those strength outlooks incorporate all available guidance when the forecast is made in early September, so they do not rely on a single or even a small number of factors. I assume that the NMME value you note includes the NASA GEOS-S2S model, which has been notably biased cold for this event, and so forecasters have been discounting its influence. The corrected NMME average values without the NASA model from early September are closer to -1C, and they won't be updated again until early October. The latest CFSv2 forecasts certainly are interesting and are consistent with what you describe, but they curiously show a rapid cooling in September that we have yet to see. Therefore, I would not base expectations on the CFSv2 alone, but it is something that forecasters will keep an eye on. The bottom line is that a weak-to-moderate event seems most likely, and a strong La Nina is unlikely. It's too early to distinguish whether it will be weak or moderate, but hopefully we'll have a better handle on intensity in October.