I looked at that link.
On NH, I basically agree with the conclusions, as I’ve already said; Bush would have won even without Nader on the ballot. However, on FL, I find the analysis shallow and tendentious. Basically, the argument is “FL exit polls show 1% Nader>Bush and 1% Nader>Gore, so Nader helped neither side.” But that’s just looking at the poorly-rounded summary tables of the exit polls; my analysis went down to the level of raw data. And yes, I did use the nation as a whole to build my model, but that is not the same as just imputing nationwide results to Florida; I built a demographic model and then applied it to make precinct-by-precinct predictions of how FL would have gone without Nader. And even if I’d done as you suggest and used just the FL exit poll results, it would have shown Gore would have been more likely to win than Bush, though the error bars would have been wider without including the national data to cross-check the model.
I also respectfully suggest that you consider whether you’re attached to this point for reasons besides its factual status. No offense intended; I know of nobody who’s always purely rational and empirical about their politics. But for the debate we’re having right now, which is a pretty simple contrafactual on which there’s some directly-relevant evidence, we should try to stay as close to that as possible.