/sci/ - Science & Math

[Anonymous No.5116759]

Sup, /sci/. I have a question.

One of my professors said that, when rounding up, you look back two numbers from the last significant figure.

An example: the number you have is 3.99483 , and you want to round it up so that there are only 3 significant figures. According to my professor, instead of referencing the 4 next to the last 9, you reference the 8 next to the 4. So the end result would be 4.00 instead of 3.99.

Is this correct or is he a retard? I don't think I'd heard this rule mentioned before.

[EK !EKFnQG0UBo No.5116767]

>>5116759
3.99483 to 3SF is 3.99
yeh, he's a retard

[Anonymous No.5116781]

>>5116759

Unless this was a Humanities course,

I recommend you try to get your Tuition money back.

[Anonymous No.5116791]

>>5116759
It depends entirely on WHERE you're instructed to round. The tens place? Tenths? Hundredths? Thousandths?

If you round to the nearest thousandth with 3.99483, it'd be 3.995.

[fizixfag !!qVqXcqA04Pc No.5116790]

Put it this way, we only include significant figures because the ones after the significant figures can't be guaranteed to be accurate, you're introducing the inaccuracy of the latter numbers by doing this, thus defeating the point.

[EK !EKFnQG0UBo No.5116795]

>>5116791
he said 3sf as in THREE SIGNIFICANT FIGURES you illiterate fuck!

[Anonymous No.5116798]

He's right, obviously.

3.99483 ~ 3.9948
3.99483 ~ 3.995
3.995 ~ 3.???

doesn't make sense, you can't even have 3.9[10] or whatever (it'd still sort of be 4 figures anyway)

the answer is 4.00 or simply 4

[Anonymous No.5116801]

>>5116795
>One of my professors said that, when rounding up, you look back two numbers from the last significant figure.
This is the incorrect advice.

[Anonymous No.5116806]

>>5116798
in the second row it should've been 3.9948 without the 3

sorry about that typo

[Anonymous No.5116807]

>>5116790
>"introducing the inaccuracy" by including more calculated digits
>somehow improving accuracy by increasing or lowering the exact name by an arbitrary amount

[EK !EKFnQG0UBo No.5116809]

>>5116798
if the 4th digit is a 4, you dont fucking round up, moron!

[Anonymous No.5116810]

>>5116809
>implying the 8 doesn't round the 4 up to a 5

[Anonymous No.5116813]

I wouldn't expect these underemployed academics on /sci/ to agree with the professor, but often in the engineering industry, x.45 is rounded up to x+1 because x.45 rounds to x.5 which rounds to x+1. It sounds silly, but the Burj Dubai is still standing and this was likely used somewhere along the way at some point so it can't be too bad of an approximation.

Your prof probably has some real world experience. Stick with them. They are not a retard.

[Anonymous No.5116816]

>3.99483
>3 significant figures
>Round to 4 anyway
I can only imagine two possible situations where this is true
>You are studying in a really shitty institution and i recommend to change schools
>You are studying physics under a "pi is roughly 3" type of professor
Either way you should just say that he is wrong

[Anonymous No.5116821]

>>5116813
just because it's good enough doesn't mean it isn't wrong. .45 does not rond to 1.

[EK !EKFnQG0UBo No.5116819]

>>5116810
that would be to 4sf, fuckhead

[Anonymous No.5116822]

>>5116819
Which is precisely when he rounded it.

[EK !EKFnQG0UBo No.5116828]

>>5116822
>"and you want to round it up so that there are only 3 significant figures" - OP
>3 significant figures
>3

FUCKING LEARN TO READ!!

[Anonymous No.5116829]

>>5116821
Well let me ask you a question in response to that. Why do we round?

[Anonymous No.5116834]

>>5116828
Look who you were quoting moron. He was listing the different ways to round it depending on which # of sf you decide to use. You told him not to round the 4 when he came to round to 4sf.

[Anonymous No.5116835]

3.99483 rounds up to 3.9948.

Then 3.9948 that you now have rounds up to 3.995.

Then 3.995 that you now have rounds up to 4.00 because you can't just "forget" that 0.005.

You should've learned that stuff in the 4th grade, OP.

[EK !EKFnQG0UBo No.5116838]

>>5116834
OMFG! how is this unclear1!L!??!!

to 3sf it's 3.99
to 4sf it's 3.995

htis aint fucking hard!

[Anonymous No.5116842]

>>5116838
...Exactly. So why the fuck did you tell him not to round the 4 to a 5 when he was rounding 4sf?

[EK !EKFnQG0UBo No.5116846]

>>5116835
>3.99483 rounds up to 3.9948.
>rounds up
>up

no...it round DOWN because the next digit is lower than a 5!

fucking retard-o-clock all up in my /sci/
fucking hell!

[Anonymous No.5116848]

>>5116838
so you're saying that 3.995 ~ 3.99 rather than 4.00? have you come here from /lit/ or something?

[EK !EKFnQG0UBo No.5116850]

>>5116842
OP said 3sf, you stupid blind twat

[Anonymous No.5116853]

>These EK threads
>EK the scourge of /sci/

[Anonymous No.5116858]

>>5116838
If I need to restate his post >>5116798 with a little more detail:
5sf: 3.9948
4sf: 3.995
This is when you interjected with your stupid as fuck, "you don't round it moron!"
He continued with 3sf just as 3.???.

>>5116850
You weren't quoting OP. You quoted someone else who was actually using 4sf.
Actually try to quote the right person next time?

[Anonymous No.5116860]

>>5116846
Right. But the end figure is still 4.00, not 3.99.

[Anonymous No.5116862]

>>5116759
it doesn't matter the whole point of significant digits is that its the limit of precision, a difference of lower order of magnitude than the least significant digit just doesn't matter.

[EK !EKFnQG0UBo No.5116863]

>>5116848
eat shit you fucking troll asswipe
you dont round twice!!

>hurr durr, 2.446 rounds to 2.45
>hurr durr, 2.45 rounds to 2.5
>hurr durr, 2.5 rounds to 3
so 2.446 = 3, fuck yerr, i haz logicks

MORON!

[Anonymous No.5116864]

ITT: People trying to argue logically over a concept that can't be proved a priori.

Rounding is subjective for the convenience of practical application you fucks. Convention isn't truth. Rounding to 3sf has no 'correct' meaning beyond typical usage.

[EK !EKFnQG0UBo No.5116867]

>>5116858
>You weren't quoting OP. You quoted someone else who was actually using 4sf.
and like i said! OP said he's rounding to 3SF
it's not my fault that he's an illiterate prick who cant fucking obey simple instructions!!

[EK !EKFnQG0UBo No.5116870]

>>5116860
only if you round twice like a fucktard

[Anonymous No.5116873]

>>5116867
And he was showing, in detail, how to round based on different situations. It's entirely your fault for replying to him before reading and understanding his post.

[Anonymous No.5116872]

>>5116863
no, 2.446 = 3 is obviously wrong

2.446 ~ 3, however, is correct

go back to school

[Anonymous No.5116876]

>>5116870
People round twice in industry and the real world all the time to great effect. Get down from the ivory tower.

[EK !EKFnQG0UBo No.5116879]

>>5116872
which has nothing to do with anything
nice shitposting, faggot

[Anonymous No.5116881]

Here's another funny paradox we get when rounding numbers.

4.999.... is rounded to 4 because it's closer to 4 than to 5.
4.5 is closer to 5 and gets rounded to 5.

[EK !EKFnQG0UBo No.5116885]

>>5116873
he was blathering on about shit no1 asked about!
3.99483 to 3sf is 3.99, end of story, /thread!
THATS ALL OP FUCKING ASKED!

fucking aspie cunts not shutting up about shit no1 cares about!

[Anonymous No.5116888]

>>5116872

Nope.... 2.446 ~ 2

[EK !EKFnQG0UBo No.5116890]

>>5116876
Then it just makes it less accurate.
you may as well round any 4 upwards, but any 4 is closer to the lower mark than the upper, so it's retarded

[Anonymous No.5116896]

>>5116881
4.999... is 5

[Anonymous No.5116897]

ITT: an underage tripfag doesn't want to admit (s)he was wrong even though it's blatantly obvious

Your professor is a well-educated individual, OP. (S)he'll probably teach you many more interesting things so pay attention and don't miss any lectures.

And don't let /sci/ bully you into thinking otherwise.

[Anonymous No.5116900]

>>5116885
Detailing how to round is precisely the point of this thread, when OP is obviously confused about it.
You're free to stop at any time.

[Anonymous No.5116902]

>>5116879
>has nothing to do with anything
it has everything to do with the post I was replying to so take a hike

[Anonymous No.5116904]

>>5116885
>demanding that questions only be answered in one specific way
>not aspie to the max

[Anonymous No.5116907]

>>5116888
0/10

2.446 ~ 2.45
2.45 ~ 2.5
2.5 ~ 3

[EK !EKFnQG0UBo No.5116909]

>>5116897
lel, so you think 3.99483 rounded to 3sf is 4.00 then? really??

>>5116900
he asked about a specific number, to specific number of significant figures!
easiest fucking shit imaginable, and you lot are still retarded.

>You're free to stop at any time.
fine, fuck this dumb nigger thread!

[EK !EKFnQG0UBo No.5116913]

>>5116907
and like i said! rounding multiple times loses accuracy!
it's closer to 2!
fucking troll piece of shit!!
i'm out

[Anonymous No.5116915]

I'm sorry that EK ruined your thread with his sperging, OP. If he actually wanted to help you he'd do it anonymously.

[Anonymous No.5116917]

>>5116909
>get proved wrong
>get mad
why do tripfriends always do this?

[Anonymous No.5116922]

>>5116913
>rounding multiple times loses accuracy
Rounding just once loses accuracy as well. Nice try though, troll.

[EK !EKFnQG0UBo No.5116926]

>>5116922
rounding again loses MORE accuracy!

[Carl No.5116939]

For all practical purposes you want to round down. Making a difference between rounding up and rounding down is pointless. Cut off the decimals.

[Anonymous No.5116947]

>>5116926
> "should we keep on rounding up that number, sir?"
> "wait, let me check if EK's accuracy-o-meter is okay with this"
Thanks for a good laugh.

[Anonymous No.5116954]

EK is right and everyone else is trolling... what the fuck is happening

[Anonymous No.5116967]

>>5116954
my first /sci/ thread and EK looks like a fucking asshat though

[Anonymous No.5116968]

>>5116954
>implying you're not EK

[Anonymous No.5116971]

>>5116967
I'm really sorry this happened to you.

[Anonymous No.5116973]

>>5116967
He gave you the correct answer. What more do you want?

[EK !EKFnQG0UBo No.5116975]

>>5116967
<<<

[Anonymous No.5116985]

Your professor is an idiot. They're going through this in a stupid way that shows that they're considering this algorithmically rather than actually thinking about it.

EVEN IF you consider the 8 instead of just the 4, it STILL should round to 3.99. Consider the 0.--48 as a single number between 0.--00 and 0.-100, or, to make it easier, as 48 between 0 and 100. Would you round 48 to 0 or 100?

You'd round it to 0, not 100, because 48 is marginally closer to 0 than it is to 100.

Similarly, there's a greater distance from 0.--48 to 0.-100 than there is to 0.--00.

So even if you include the 8, you'd still end up rounding to 3.99. But your professor isn't thinking about the actual reality of this, he's just churning through and comparing numbers in a stupid way.

[Anonymous No.5116988]

OP here, I've been afk for a while. My thanks go to everyone who answered (even to EK who seems to be losing it more and more every day since he started using a tripcode), but since these are engineering-related classes, I'll go with what >>5116813 said.

[Anonymous No.5116993]

Rounding up on 0.5 is only a convention, anyway. It produces an upward bias.

To remove this bias when it's significantly undesirable, there are other conventions, such as "round to even": 0.5 rounds to 0, 1.5 rounds to 2, 2.5 rounds to 2, 3.5 rounds to 4, etc.

[EK !EKFnQG0UBo No.5116997]

>>5116988
>even to EK who seems to be losing it more and more every day since he started using a tripcode
ive used a tripcode since forever, and i aint 'losing' anything!

the point in rounding is to make it simpler when you dont need an exact answer, but you're still supposed to round it the closest way to keep with accuracy.

i've no idea why >813's proffessor said a 0.45 should go up to a 1, rather than down, but w/e

[Anonymous No.5116999]

mfw as a mathfag I won't ever have to deal with this garbage.

[EK !EKFnQG0UBo No.5117000]

>>5116993
the point was that after the 0.5 there MIGHT be other digits, and if there are, then it DEFINITELY rounds upwards
if it's exactly 0.5, then yeh, it's equally close to both, but the original point was that there might be more to it (it might be a rounding itself,) so you'd go upwards

[Anonymous No.5117001]

>>5116997
>i've no idea why >813's proffessor said a 0.45 should go up to a 1, rather than down, but w/e
Because they're using a shitty way of rounding. They're recursively applying the "if it's 0-4 round down/5-9 round up" rule in loops from the back of the number but only on single numbers at at ime instead of considering it as a generalizable rule.

It's what happens when you learn math and science by rote memorization instead of stopping to think about why you're doing it. Sometime in the past, he learned the rounding rule and how to apply it and stopped right there.

[Anonymous No.5117002]

>>5116798
0.4444444444445
rounds to 1 apparently.

[EK !EKFnQG0UBo No.5117009]

>>5117002
that was my point in: >>5116863

rounding multiple times loses accuracy

[Anonymous No.5117022]

>>5117009
>loses accuracy
here we go again

don't you trolls ever give up?

>>5116922 and >>5116947 already addressed that btw

[EK !EKFnQG0UBo No.5117026]

>>5117022
see >>5116926
you cunt

[Anonymous No.5117039]

>>5117026
And how to tell when the accuracy loss becomes such that it can no longer be considered acceptable, troll? Please cite an authoritative source in your answer.

[EK !EKFnQG0UBo No.5117043]

>>5117039
eat shit
if you dont know how to round properly to the CLOSEST FUCKING THING then gtfo, nigga

[Anonymous No.5117044]

>>5117000
My point is, people shouldn't be arguing as if rounding up on 0.5 makes some kind of inherent sense. It doesn't. It's just a convention.

For different sorts of measurements, even rounding to the closest digit isn't always best. Sometimes, you get better results if you bias it one way or the other, particularly if something like a logarithmic scale is involved.

You can find a depth even in subjects considered elementary, like arithmetic.

[EK !EKFnQG0UBo No.5117052]

>>5117044
>For different sorts of measurements, even rounding to the closest digit isn't always best. Sometimes, you get better results if you bias it one way or the other
ah, like jews always rounding the bill upwards?

[Anonymous No.5117073]

>>5117043
>It's the right way of doing things because I rudely said so!
That's tripfags for you, folks.

[Anonymous No.5117074]

mathfag here, isn't this effectively the same as rounding to 0.4444.. (or 4/9) as opposed to rounding to 1/2? Why would you do this?

[Anonymous No.5117082]

This is plain stupidity.

What this moron suggests is that if you want to round
0.445 to the nearest tenth, you'd get .5

[Anonymous No.5117100]

>>5117052
something like that

[Anonymous No.5117119]

>>5117082
>stupidity
>moron
Just because you don't understand it doesn't make it invalid.

[Anonymous No.5117139]

>>5117074

I drew you retards a picture.

[Anonymous No.5117143]

your professor is obviously an idiot

[Anonymous No.5117146]

>>5117119
I understand it, and I think it's stupid.

[Anonymous No.5117152]

>>5117146
If you truly understood it, you wouldn't think of it as stupid.

[Anonymous No.5117158]

>>5117152
it is retarded, refer to
>>5117139

[Anonymous No.5117176]

>>5117158
it's not, refer to
>>5116835

[Anonymous No.5117217]

>>5116835
this is the same thing.

3.9948 is greater than 3.99 + (1/100)(4/9) which means it rounds up to 3.99 + 1/100 = 4

The difference is that you're rounding to 4/9ths instead of 1/2. Which is totally retarded.

[Anonymous No.5117256]

you don't round more than once. you only round once at the end of your calculation. you also choose the number of significant figures by your least precise result.

for example:
4.4462g of salt is dissolved in 103ml of water.
determination of concentration the most significant figures you can use is 3.

[Anonymous No.5117500]

bump