End The Whing-O-Rama, Good Advice From Bojo?

[ToraToraTora]

http://news.sky.com/story/snap-out-of-donald-trump-doom-and-gloom-says-boris-johnson-10652841

[jim360]

And I "bleated on about science" because you and JD brought it up, apparently gleeful that I'd made such a glaring mistake, when I haven't made any errors at all. Except, perhaps, trusting people to pay attention to the substance of an argument, rather than getting distracted by minutiae.

[Zacs-Master]

You're arrogance in claiming you haven't made a mistake is astounding. It undermines your whole reasoning.

[bhg481]

Jim, like many other scientists, is used to "back of envelope" calculations, where approximate values are "good enough" to reach a conclusion. He is trying to explain that, in this case, there is no significant difference between 3 and 4 times; the relevance is that both figures are significantly different from the value 1.

[jim360]

Not really arrogance if it's accurate though.

None of the figures I have provided are wrong in any meaningful sense. They are rounded, they are estimates, and they are meant to support a broader point. If you are so bogged down by a 25% difference (which anyway is misleading on *your* part, because I never said it was exactly 4 or exactly 3), I can't do anything about that.

The point is that the electoral college provides for serious imbalances between individual voters depending on the state they live in. The precise size of that imbalance is neither here nor there, as long as it exists in some measure. I attempted to give an estimate for the size of that measure; as seen, depending on the assumptions you make, it works out to be around 3, or around 4, at its most extreme.

These imbalances do somewhat cancel out in practice, because the "unfair" states where voters have much more power than they should are roughly split. Of the ten worst offending states, Trump got five and Clinton got five. Meanwhile, although Clinton has won the popular vote, she appears to have done so by a single percentage point, which means that in the national tally the voter imbalance ends up being far smaller than 3 (or 4). Again, though, it still exists and should not. I can't be clearer than that. I would love to understand why anyone could defend the electoral college, as birdie did (and Naomi in supporting his opinion), when it creates such imbalances that allow the will of the people as measured by popular vote to be ignored on, so far, four occasions in US electoral history (five, if you count 1824, which is a little more contentious).

[jim360]

Thank you, bhg.

[Zacs-Master]

Back of envelope calculations where the basic data (population rather than voters) is seriously flawed, leading to a 25% error in the result. Mmm.

[jim360]

It's not all that flawed, actually. The Electoral College is determined based on state populations, not voting populations -- so it's a fair starting point for any comparison. The discrepancy emerges because each state is given two EC votes (two represent two senators) to start with.

A 25% error, at this level of working, is simply not that big a deal.

[jim360]

To address your point in another way... well, again, you seem to be missing the point of what a back-of-envelope (BoE) calculation even is, and what anyone who does one is looking for. Almost by definition, the data that goes into a first BoE calculation has to be flawed in some way: if it weren't flawed, it wouldn't be a BoE calculation. All it has to be is "good enough" to land close-ish to the exact answer. Usually, this means being about the right size. What that means varies from case to case, but here it seems to me that, as I was looking for an answer markedly different from one, anything between, say about 3 and 5 was basically the same size -- or, put a slightly less convenient-looking way, if I found a ratio of "about" 1 or 2, say, then I'd probably have to either find more precise figures or try a different pair of states.

Since I did get an answer very different from 1, so as to be "good enough"; since the electoral college is based on state population not voters; and since the errors these two assumptions were bound to introduce were, likely to cancel out to a large extent (they often do in ratios), I was happy with the figure. I would have tolerated an error of probably 50% down and it would still have made the point I wanted to make. Turns out the error was less than this.

I would call 25-ish% error, based on two minutes' work, a pretty decent success.

[jim360]

OK, I'll admit it. I did make a mistake after all.

In the second case, I just went and checked over the figures more thoroughly this time, and while I was right that Alaska had the lowest number of votes for the victor per electoral college vote, I was wrong to choose Maine as the most extreme comparison. It looked plausible, at a first glance, looking over the data by eye.

I've just gone back and computed the ratio (votes for winner)/(electoral college votes) for each state, and in fact the state which had the most votes cast for the winner per electoral college vote is Massachusetts. The ratio between that and Alaska?

4.11.

[Zacs-Master]

Jesus.

[jim360]

Well, you did ask for more precise figures, presumably. I've gone and done them for you. Make up your mind what you want.

[naomi24]

I'm rather pleased I dropped out of this conversation some hours ago.

[Zacs-Master]

I wanted an admission you were wrong which, after a lot of smokescreening, irrelevancies and general cobblers you gave me. The 'jesus' was in reference to it being couched in a load of irrelevant waffle which just makes you look even dafter than you did when I knew you were wrong a couple of hours ago.

[Zacs-Master]

We're all grateful Naomi.

[jim360]

I *was* wrong, but not in the way you seem to think. I wasn't wrong to produce a ballpark estimate (which turns out to be have been far more accurate than I had anticipated), because that's what you are supposed to do. I wasn't wrong to try to get more accurate figures, because that's what you are supposed to do next. I *was* wrong to compare the two states I did. It's a very minor mistake, and doesn't at all affect any of the point I was making. In the end, it seems that my "seriously flawed" initial estimate was, what, less than 3% out?

Now, perhaps you would care to address the actual point, which is the travesty that is such a serious discrepancy in voter power across the US, rather than imagined flaws in estimation techniques?

[Zacs-Master]

You just can't help yourself can you.

[jim360]

Of course I can't. You are accusing me of being wrong in a way I don't recognise to even remotely accurate. So I'm going to deny it and refute it as far as I possibly can.

Of what, now, are you accusing me?

And, again, can't you just answer the actual question?

[naomi24]

Zacs-Master //We're all grateful Naomi. //

I don't doubt you are at least.

[Jackdaw33]

Must be a quiet week in the world of quantum physics. Isn't there a rogue quark waiting to be discovered which might be more worthy of Jim's attention than the US political system?

[Peter Pedant]

// You're arrogance in claiming you haven't made a mistake is astounding. It undermines your whole reasoning.//

Hahaha o God Jim did you read that ?

ZM - Jim's major is Maff n Logic
and take it from me - they dont say 1+1 =3 less than 2% of the time
instead they er never say it


advanced maff teachers ( like Jim ) really really dont make mistakes ....
not even sometimes -

and the logical mistakes and non sequiturs of others ( an everyday occurrence on AB in every thread ) really do glare at them like the red hot coals of a rabid dog ........

Jim I still think you should retrain in Law ....