Silicon Investor (SI) -- The First Internet Community

STOCKTALK

We've detected that you're using an ad content blocking browser plug-in or feature. Ads provide a critical source of revenue to the continued operation of Silicon Investor. We ask that you disable ad blocking while on Silicon Investor in the best interests of our community. If you are not using an ad blocker but are still receiving this message, make sure your browser's tracking protection is set to the 'standard' level.

Politics : New FADG. -- Ignore unavailable to you. Want to Upgrade?

To: Nadine Carroll who wrote (1611)	6/11/2007 11:12:32 AM
From: neolib	Read Replies (1) \| Respond to of 4152

Um, the point Freeman Dyson was making was that of course they do. They have been munged to fit the data. This alone makes them scientifically unreliable for predictive purposes. That is where he does not understand the difference between a model and a curve fit. His entire discussion, and most of his reader's comprehension is that somehow the "model" is more like a curve which has been fit to the data. You can always fit some sort of curve (splines for example) to a set of data (assuming the data is reasonable). This is not what a dynamic system model is. Clearly someone could curve fit to the data, then pick any random forward projecting curve they want. In that sense they could have a "model" which fits the historical data, but gives their desired outcome going forward. I did not think fit to suggest that before, because I assumed you understood this is not valid. Your reply here leaves me realising my mistake. There are two things a real model must do: 1) Produce outputs which fit known results when given known inputs, i.e. validation, and 2) the model equations must also match known science. #2 gets rid of a model which is simply a curve not based on science. You do realize that this statement is equally valid interpreted as a) the current outputs are the only valid outputs, and b) the current outputs are unreliable, but improving them is beyond our current scientific know-how. Nonsense. There are plenty of models out there and they do differ. They just get discarded if they can't do a reasonable job on a historical data set. Why would anyone accept a model which cannot valid on the existing data set? The scientists of 100 years ago discovered many things for which there was scientific evidence, and have held up to later discoveries. Like special relativity. Eugenics did not hold up not because its proponents were ignorant, but because they jumped their conclusions in an unrigorous way ahead of available data, based on what they truly hypothesized to be correct. It was then advanced further by the various do-gooders and politicians. It should stand as a warning not to politicize science and not to advance its conclusions beyond the availability of sound scientific evidence. You fail to understand which sciences were more advanced 100 years ago. Math, physics, and chemistry were doing significantly better 100 years ago than biology/medicine compared to current knowledge. No surprises there. BTW, although I have not read much about eugenics, I suspect there was a good deal more "correct" with it than people care to admit. One certainly can make observations about human reproduction and evolution, the consequences of which we might not like. This does not mean the observations are wrong. Clearly pop evolution can lead to simple racism, which may well be incorrect. But modern human society has indeed intruded significantly into the dynamics of our own species evolution, and the effects of this are hardly understood at this time. The welfare state has a very very short run compared to our species history, so stay tuned. My own POV is that intelligence can in fact allow you to safely intrude on evolution and in a way which will not implode long term and is also kind and moral. But very little intelligent thought is given to this, as most people don't believe those wicked scientist who say humans evolved.