[Cascade]

barmar, on Jul 8 2008, 06:39 PM, said:

Quoting bridge percentages to 4 significant figures is like a weatherman predicting a 49.27% chance of rain. We don't even believe he can be accurate when he rounds to the nearest 10 percentage points, so extra precision is meaningless. And even if he IS that accurate, what can we do with all that extra precision? You're either going to take an umbrella or not -- does a hundredth of a percentage make a difference in that decision?

Do you take an umbrella everyday?

If not then at some point (range) over a range of 1/100 of a percent it does matter.

[effervesce]

The problem is is that if someone says a number as xx.xxx, it technically means that they know the answer to that number of decimal places of accuracy - for example, if I said 10.00mL as instead of 10mL, 10.00mL means (at worst) 10.00mL plus or minus 0.05mL, while 10mL means 10mL plus or minus 5mL (at worst).

Hence, when the kibber said 55.555%, he's technically saying the probability is 55.555% plus or minus 0.005%, which is ridiculous.

For example, if a person said they walked half a kilometer - did they travel 500.000m? No, they meant 500m, plus or minus 100m (200m?) or so.

[TimG]

effervesce, on Jul 8 2008, 05:37 PM, said:

Hence, when the kibber said 55.555%, he's technically saying the probability is 55.555% plus or minus 0.0005%, which is ridiculous.

Would it have been ridiculous if he had said 5/9 instead of 55.555%?

[655321]

TimG, on Jul 8 2008, 07:36 PM, said:

effervesce, on Jul 8 2008, 05:37 PM, said:

Hence, when the kibber said 55.555%, he's technically saying the probability is 55.555% plus or minus 0.0005%, which is ridiculous.

Would it have been ridiculous if he had said 5/9 instead of 55.555%?

TimG, so many of your posts are of this nature. I don't see the point.

[TimG]

Sorry, I guess I did already ask this question, perhaps a private reply to effervesce would have been in order if I felt it necessary to repeat.

[nige1]

matmat, on Jul 5 2008, 08:29 PM, said:

I was kibbing a team match today. At one point during a discussion of a possibility of a particular holding by a 1NT opener a kib claimed the percentage to be 55.555%. They were being totally serious about this.

My kneejerk reaction to this is always that giving percentages to anything smaller than a whole percentage shows a lack of understanding of the game and is, well, stupid.

What do you think?

5/9 may be more accurate but I settle happily for 56% or "better than evens" although it could make a difference to a book-maker

[jtfanclub]

Interesting. If somebody says that this is a "55.55%" or "5/9", this is useful to me, because then I can frequently recreate the method he used. If the person says "slightly over 50%", or even "55%", how do I determine if the person is right or not without starting from scratch?

And yes, the odds really are that precise. This isn't chemistry, or statistics. It's very common for the odds of a finesse to be 50.0000000000000000000000%, and I left out a whole lot of zeros. Bridge is more akin to accounting. And in accounting, just because you're working in sums of millions of dollars doesn't mean that you just round off to the nearest thousand. It's the smaller parts that tell you when a mistake has been made. You're never off by a penny. You're off by a whole lot of dollars and a penny.

If a person were to say that the odds were 54.545%, and I can see that there's no way that there's 11 or any multiple of 11 possibilities, I know he's made a mistake. It's not that I care about the difference between 54.X% and 55.X%. If he's got the denominator wrong, he's probably off a lot. It's more likely that the right answer is 6/9 or 4/9 than 5/9.

So yeah, I want those pennies!

[effervesce]

TimG, on Jul 8 2008, 07:36 PM, said:

effervesce, on Jul 8 2008, 05:37 PM, said:

Hence, when the kibber said 55.555%, he's technically saying the probability is 55.555% plus or minus 0.0005%, which is ridiculous.

Would it have been ridiculous if he had said 5/9 instead of 55.555%?

Ah sorry - I didn't realise that the actual real probability was 5/9 - then in fact, 55.556% is better than 56%, but 5/9 better than either.

[barmar]

Assuming the actual probability is 5/9, then they could say 56%, 55.6%, 55.56%, 55.556%, and so on. When they say 55.56% it suggests that they know it more precisely than 55.6%, but not as precisely as 55.556%. That's the problem with expressing things like this as decimals -- it implies a specific level of precision. In particular, when they say 55.55%, rather than 55.56%, it means that it's closer to 55.55% than 55.56%, which is not correct for 5/9.

They could, if they wished, say 55.55... -- the elipsis indicates a repeating decimal, which is what you get when you try to convert 5/9 to decimal. But if they're going to do that, why not just use the fraction?

[jtfanclub]

barmar, on Jul 8 2008, 10:38 PM, said:

But if they're going to do that, why not just use the fraction?

I would assume the person put a number of fractions into a calculator, the others canceled out, and he didn't notice when he typed that it equalled a fraction.

[Rossoneri]

jtfanclub, on Jul 9 2008, 11:53 AM, said:

barmar, on Jul 8 2008, 10:38 PM, said:

But if they're going to do that, why not just use the fraction?

I would assume the person put a number of fractions into a calculator, the others canceled out, and he didn't notice when he typed that it equalled a fraction.

That's possible. Experience has proved that most people don't really recognize what fractions correspond to what percentages unless it's like 50% or 25% etc.