I'm not defending Clinton in this discussion. I would have the same analysis if we were talking about Trump. It would be the same if we were saying getting a credit card number ending in 2222 2222 is probably not random because the chances are 1 in 100 million. I would have to point out that any eight digits of the same number would look suspicious, e.g. 3333 3333. So would 1234 5678 or 8765 4321. We have to add up the probabilities of all the cases that would appear anomalous.

it would make a great movie -- who can we call? -- j
.

The Notebooks of LL were great, and full of good practical advice (though I have managed to get this far in life without ever learning how to diaper a baby!!! - and I intend to keep that lacuna in my knowledge base as long as possible). It would be great to see TMIAHM made into a movie.

Jan

I understand what you're saying, but it is a complete distraction from the original question - which ignored Bernie Sanders' role entirely. No one was concerned about Bernie Sanders cheating. No one really thought Hillary would need to cheat before the fact. What we are analyzing is one simple question: what are the odds Hillary - and only Hillary - cheated?

Anomaly hunting is done by those attempting to justify that an outcome - while being unusual - has enough precedent to be considered legitimate. If you want to try to apologize for Hillary's cheating (which is what this is seen as) you can certainly go down that road if you choose. I find it unlikely, however, that - in light of Hillary's past behavior - you are likely to find many on this forum willing to join you.

Now that is an interesting point! I do remember LL saying that, I think it gets repeated several times in several books. I hadn't even remembered Clarke's laws, thanks.

Same old stuff. She gets away with high crimes and misdemeanors.

Cheating? What cheating? There are several Democrat members of the Coin family in Iowa: Henry Coin, Jane Coin, and more. Six of the Coins are in different precincts. They are all for Bernie, but the rules say the result can be determined by flipping a Coin. All six of the were flipped by Clinton cash. All in accordance with the rules.

I didn't get past the time enough for love era. . the
excerpts from LL's notebooks got me. . isn't it wonderful
how they look like orchids? . lovely! . / little girls and
butterflies need no excuse. -- j
.

Cummon now. Clarke was awesome too. And his 3 Laws are all worth remembering.

Yes, Heinlein was also my favorite for a long time. I fell off the back of the truck when I read some of his later novels (and still have not read many of them).

Jan

thank you for the info, Jan; I preferred Heinlein! -- j
.

“The strict probability of the outcome of six straight beneficial binary decisions is 1/2^6 = 1/64. You are arguing that the probability is only 1/32”
First, I’m very sorry for sounding like I’m going off. I’m just interested in this type of thing b/c this issue comes up so often, and I think I see it through the weird lens of electronic communications theory.

It’s a straightforward math problem if someone asks before an experiment “what’s the chance of getting six heads from six tosses.”: 1 in 64. My claim is that if we’re looking for a non-random element to the tosses (e.g. cheating) in the data, we have to consider all the cases we might consider anomalous. I consider either candidate winning all the tosses to be equally anomalous. These two cases make us wonder, “What are the chances that candidate X won all six tosses?” The probability of getting a result that makes us wonder that is 2 in 64, or 1 in 32.

I think the fancy name for this is post hoc anomaly hunting. People can take it to an extreme and find post hoc anomalies in any permutation of outcomes. They could check to see if the tosses lined up with who was mayor in those areas, who was county executive, who was the Congressman, whether they were more urban or rural, until they come upon an anomaly. Then they ask, “What are the chances that in every area that had a Democrat county executive the toss went for Clinton, and it went for Sanders in all the others? It’s straight math: 1 in 64.” Obviously you and I are not doing that; it’s just an extreme example of anomaly hunting.

I would think either candidate winning all the tosses sounds anomalous, so when asking myself “what are chances” I have to consider both scenarios.

I'm doing a mathematical/logical analysis on how you came up with your 1/32 claim. I don't what else you're going off on and I'm not dealing with or incorporating statistical anomalies at all.

The strict probability of the outcome of six straight beneficial binary decisions is 1/2^6 = 1/64. You are arguing that the probability is only 1/32. The only way you could come to that answer was by twisting the original problem statement to apply to either of two candidates and aggregating/summing the individual odds and then presenting that as the real odds of the outcome - or just by being bad at permutations. I'm pointing out that your answer is to a different question than the one originally asked, which is explicitly "What are the odds that Hillary Clinton was benefited from the outcome of six straight binary decisions (coin tosses)." It's a straight math problem.

Now if you want to argue that despite it being a statistical anomaly it isn't prima facie evidence of fraud, you can make that argument. You have a 1/64 chance of being right, but you can make the argument. I'm looking at it and going - with those odds, Hillary should be down at the track.

Here is the market watch story
http://www.marketwatch.com/story/coin...

Here is the story from NPR
http://www.npr.org/2016/02/02/4652682...

I have seen it reported that she actually won 6 out of ten. We are talking about the press so I am reticent to say one way or the other. :-)

"not the aggregate possibility of two future outcomes."

This is a logical fallacy. If we're using the apparent odd coincidence as possible evidence the toss was not random, we must consider all possible odd coincidences*. Post hoc anomalies (I think that's the right phrase for this) like this crop up everywhere. They make people believe in things like ESP. "What are the chances that I would dream a car accident and then have a car accident shortly after?" they ask. But they need to ask what are the chances there would be any possible dream followed up by any possible similar occurence. If you go hunting for any anomalies in the past, you will find them, but they don't serve as evidence for anything.

BTW, this does not mean the coin toss was fair. You may have other arguments based on Clinton's past behavior to suggest she would be willing and capable of rigging a coin toss. My claim is the anomaly itself is not evidence for anything.

Here is how we do it in Washington state...

Toss #1 - Sanders
Toss #2 - Sanders
Toss #3 - Clinton...Clinton wins!

Why not...that's how a previous Gubernatorial race, between Democratic and Republican candidates was handled.

So, did she win 6 out of 6 tosses or 6 out of 10 tosses? The former raises an eyebrow, the latter is quite believable.

You are using "either" in an inclusive manner which equates to the statement the combined odds of Clinton winning (1/64) PLUS the odds of Sanders winning (1/64) = 2/64 = 1/32. Your statement is implying the chance of future shenanigans are equivalent on either side.

The actual question, however, is applicable only to the one candidate - in this case Clinton - since what we are describing was the actual outcome - not the aggregate possibility of two future outcomes.

1) statistically this would be interesting, but given the public evidence that they were in fact coin tosses witnessed by voting observers, only interesting.
2) I have seen it reported that there were more like 10 coin tosses of which Mrs. Clinton won 6. less interesting.
3) finally we are talking about 6 of 11,065 delegates. If this is a scam it become another reason not to vote for her, not because she is gaming the system, but because she is so bad at it.

For myself I do not need to makeup reasons to not support Mrs. Clinton.

Two things...
First, since Clinton is involved, I'd venture that the odds of something shady happening are more like 1:1.
Second, a statistician expert where I worked calculated how likely the first Viet Nam Draft Number pick (remember the ping-pong balls?) was actually random.... something like 1:67,000.
Trust no one.

They always seem to... :)