Bankers Rounding
-
> ... IBM, IEEE, Microsoft, about every signal processor manufacturer in the world, the PLC manufacturing industry, and just about anyone else who matters.
Once again you manage bring up something that has nothing to do with the subject at hand.
_____________
Code for TallyGenerator -
Okay, here is your task:
You have collected 0.50 in taxes from the sale of 1 or more items (doesn't matter how many). You have to distribute these funds to three seperate taxing authorities, to the nearest hundreth, your distribution can not exceed the total collected, you must use the same rounding method for all calculations. The distribution, to the thousandsth is A (Federal) - .250, B (State) - .125, C (Local)- .125. Any funds left over go to the Local Tax Authority.
-
"Once again you manage bring up something that has nothing to do with the subject at hand."
Or now you're trying to prove that Bankers Rounding is invalid?
Don't bother, I know that.
Normal rounding brings:
0.13
0.06
0.06
-----
0.25
_____________
Code for TallyGenerator -
I'm not trying to prove it is invalid. All rounding methods have some issue some where. Will you agree with that statement?
I am not proposing that Bankers Rounding be used in all situations, either.
I used it is a specific application. Its one draw back, it sometimes came under the actual amount collected. It NEVER was over, unlike using the standard round function, which could go EITHER way. Using the Banker Round function reduced programming complexity in this case resulting in having to handle only one exception, and given a specific business rule, how to handle it. In addition, the number of exceptions handled using this was much lower than had we developed processes to handle exceptions for both under distributions and over distributions of tax amounts.
And you think this is wrong? Sorry, it worked, was approved by the accountants, and the auditors, and for this company is a part of their business practice. Once again, it isn't always cut and dry. You have to be willing to make compromises.
-
Also, your numbers were wrong: go back and read. You collected 0.50 in taxes, and you need to distribute these amount to three taxing authorities: .250, .125, .125. .25 + .125 + .125 = 0.50, the amount collected. Normal rounding to 2 decimal digits would result in .25 + .13 + .13 = 0.51. This is a violation of the constraint that the distribution exceeded the amount collected. Using the BRound funtion, however would result in: .25 + .12 + .12 = 0.49. This is not a violation of the constraint, but does result in an under distribution of taxes collected. This is then handled by the exception rule, that the extra penny (in this case) be added to the local taxes collected. This rule could have been to Federal or State, or County, but it is a rule determined, not by programmers, but by business decision makers.
-
> Sorry, it worked,
Sorry, it did not work.
Everyone can see it.
> was approved by the accountants, and the auditors, and for this company is a part of their business practice.
It does not matter how many people approved geocentric model and how much every day experience and most of people observations (including very famous and respected scientists) were convincing it's right.
It's wrong.
_____________
Code for TallyGenerator -
Sorry, again, you failed to answer all my questions.
-
And wether the earth is the center of the universe or not has no bearing on the use or validity of the Bankers Round function. They are not related in any way, shape, or form. It is like comparing rocks to fruits. They have nothing in common, except that the fruits grow on plants in the ground.
-
They have at least something in common:
- they based on wrong assumptions;
- they were proven by some people (by you in case of BR) using big names of big authorities. Nothing in common with finding the truth.
_____________
Code for TallyGenerator -
Know what, you win for all the wrong reasons. I am not going to continue this inane argument with you any longer. You aren't even willing to consider that the function as any application in any application whatsoever. If that is not arragance, I don't know what is.
-
And I don't what your B.S. degree is worth but you perception of business practices is just wrong.
In tax reports if intermediate results must be included in report
TotalTax should not be equal to round(TotalSalesAmt * (TaxRate1 + TaxRate2 + TaxRate3), 2)
It's equal to (TaxAmt1 + TaxAmt2 + TaxAmt3).
No matter which rounding was used for calculating TaxAmt's.
I can see it every day in Australian invoices containing GST per line. In many cases TotalGST 10% * TotalAmount, but TotalGST = SUM(LineGST), where LineGST = ROUND(LineAmount * 10%, 2).
In all tax reports all further calculations must be based on rounded intermediate result, if there is such.
_____________
Code for TallyGenerator -
100 replies.
Whoo-hoo!
And if to consider that this topic was started as an extension to another one, also quite long...
_____________
Code for TallyGenerator -
I took your test case and proved you wrong and you still claim to be correct. As my Dad used to say "If the hole is getting too deep, stop digging."
-
David, there was a question buried in tons of spam:
What would you name a "midnight event"?
What does it mean for you: this happened at 00:00?
_____________
Code for TallyGenerator -
"Does the name "Giordano Bruno" say to you anything?"
Sure it does. Sometimes, the little guy is right and everyone else is wrong, but in this particular case, the rest of the world is Giordano, while you are the Catholic Church, so you can stop trying to light a fire under your feet. The "TimeCube guy" also says something to me. Sometimes the little guy who thinks they are right, and that everyone else is out to get them, is in fact a nutcase. I'll leave it to each individual to determine which one better represents you in this particular area.
Look, I'm an optimistic guy, and think that everyone can eventually learn something, as long as they are open to the fact that they don't know everything already, so I'll try this one more time.
This rounding method is meant to serve a single, very simple purpose. It is intended to reduce the bias caused by traditional rounding over a very large, randomly distributed sample set. That is all. You have yet to provide any evidence (and I do mean any) that it doesn't do so. Feel free to correct that. You already tried to show it didn't work once, and that backfired on you rather badly Giordy, but I'm patient, and understand that wannabe martyrs sometimes need a few shots at fame, so feel free to try and come up with a new appropriate test. Hopefully, this time the test won't serve as a drilling rig in your hole digging process, and we'll all crawl back into the holes from whence we came. With bated breath, and whispring humblenesse, I await your demonstration of rounding prowess.
Viewing 15 posts - 91 through 105 (of 373 total)
You must be logged in to reply to this topic. Login to reply