I don't see why crossing the x axis should have any impact on the rounding convention used.
logically, the minus shouldn't make any difference. -0.5 rounds up to 0, -0.51 rounds down to -1
Shirley?
I'd have said the exact opposite... 0.5 is closer to 1... so -0.5 is closer to -1. Surely you have to be symmetrical around the x=0 axis.
Only if there's something magically special about 0. Being consistent if you round up then 0 is up on -0.5
If you view a set of numbers to two decimal places then:
4 is the range 3.50 to 4.49 (100 numbers to 2 decimal places) 3 is the range 2.50 to 3.49 (100 numbers to 2 decimal places) 2 is the range 1.50 to 2.49 (100 numbers to 2 decimal places) 1 is the range 0.50 to 1.49 (100 numbers to 2 decimal places)
So logically and consistently 0 should be the range -0.50 to 0.49 (100 numbers to 2 decimal places)
Otherwise 0 contains a range of just 99 numbers to 2 decimal places and I see no reason why it should. The core of mathematics is to be consistent in how you treat your numbers.
I think the standards are rounding to even (accountants' rounding), or rounding away from zero. Those are the options in the programming language I use, for example. Excel also apparently rounds away from zero (though I'm on my phone, so I can't check)
But really, this isn't maths, it's just about choosing between conventions. Pick whichever one works for your situation. There's no right and wrong, just useful and useless.
Apologies for the negativity but.... I really don't think polling/election predictions/psephology is something the lords should consider official business. Leave it to the electoral commission/BPC etc.
There's an insane amount of really difficult, dull but essential brexit scrutiny for the lords to be getting on with. They should be working overtime on economic impact assessments and whatnot.
Hope the day goes well, anyway. I don't blame you for accepting the invitation!
The couple of times I've seen you on TV you came across pretty well.
It's of political interest because of the view that polls should be banned in the final week(s), as in many other countries.
I don't see why crossing the x axis should have any impact on the rounding convention used.
logically, the minus shouldn't make any difference. -0.5 rounds up to 0, -0.51 rounds down to -1
Shirley?
I'm unsure that there's any convention over this wrt rounding up, down, away from zero etc. But as an engineer, there's a danger in saying '0' when '-0.5' is meant. Zero is taken too readily to mean nothing is present, when there is actually something present. Likewise, there might be a special meaning in the fact it is negative. Hence, in the absence of any other rules, I would round down to -1. Likewise, with +0.5 I would round up to +1.
What if it was - 0.4?
If the mere fact that it is negative is important, then you cannot have a rounding to integer rule. Round to a tenth, not integer. The round rule interval has to match the sensitivity of the system.
I don't see why crossing the x axis should have any impact on the rounding convention used.
logically, the minus shouldn't make any difference. -0.5 rounds up to 0, -0.51 rounds down to -1
Shirley?
I'm unsure that there's any convention over this wrt rounding up, down, away from zero etc. But as an engineer, there's a danger in saying '0' when '-0.5' is meant. Zero is taken too readily to mean nothing is present, when there is actually something present. Likewise, there might be a special meaning in the fact it is negative. Hence, in the absence of any other rules, I would round down to -1. Likewise, with +0.5 I would round up to +1.
What if it was - 0.4?
If there were no rules specified, I would be tempted in certain circumstances to make it -1, as long as I clearly specified that was what I was doing.
Say I have a set of outputs from a machine:
0.32 0.45 0.57 1.03 0.03 -0.40 0.12
The -0.40 value is interesting, especially in cases where it is important that the numbers do not go negative, or where negative cases are special. If you take the round-to-zero approach, then the one potentially-strange value in there would get lost as it is rounded up to zero.
But again, it depends on what you are using the values for.
Comments
But really, this isn't maths, it's just about choosing between conventions. Pick whichever one works for your situation. There's no right and wrong, just useful and useless.
Going with the colour/fruit match, maybe damsons.
NEW THREAD
new thread
Say I have a set of outputs from a machine:
0.32
0.45
0.57
1.03
0.03
-0.40
0.12
The -0.40 value is interesting, especially in cases where it is important that the numbers do not go negative, or where negative cases are special. If you take the round-to-zero approach, then the one potentially-strange value in there would get lost as it is rounded up to zero.
But again, it depends on what you are using the values for.