Ummm... I have a question about the discussion of when we should pay attention to data or prediction. Shouldn't we pay prime attention to the present when considering new record bad conditions?
It seems to me that predictions are nice to have, but it's not like they're actual data, and it's also not like we could avert anything by knowing the prediction a few days in advance. And predictions that fail to manifest seem to harm credibility.
AmbiValent, you have a strong point about in your urging us to pay attention to actual data, in my view, and that is what the area and extent thread is all about, and the PIOMAS update thread to a lesser degree. I have argued the point myself. I think that most posters on this thread also pay close attention to data, but ... it is also interesting and instructive and just plain fun to look ahead. Scientists don't just analyze data retroactively, they test their understanding by making predictions and seeing how they work out. You are right, however, in saying that sometimes people get carried away and I agree this can trivialize this thread.
Weather forecasts up to three or five days into the future are often reliable, and as we see those unfurl in real time I find I learn a lot. As magna recently reminded us, and as Sterks does on a regular (and needed) basis, forecasts more than 5 days into the future are not so reliable, and if they are seized upon as near certainties then posters are often proven incorrect and this leads to the 'well it wasn't as bad as we thought' revisionism, or just plain unembarrassed radio silence. The same can apply when a sudden event, like the recent ice fragmentation in the ESS, lead some of us to think 'Oh boy, this is it!' and then more experienced hands remind us that in fact the ice is often more resilient than we think. I have been guilty of this myself, and am gradually learning and gaining experience.
So, agreed on the centrality of data. But I think it is not just fine, but scientifically healthy to make well-reasoned, thoughtfully calibrated predictions. And if some of those don't turn out to be 'true', that is the way it is with predictions and we are still actively learning ...