Zugg Software :: View topic

Posted: Sat Sep 24, 2005 8:08 am

Most of the scripts I write are designed to track statistics and averages. One of the most common formulae in all of my scripts is:

Apprentice Joined: 20 Sep 2005 Posts: 135

#VAR qbonusta {%concat( %eval( @qbonust/@qcompleted), %if( %len( %eval( (@qbonust*100/@qcompleted)/@qcompleted))<2, %insert( ".0", %eval( (@qbonust*100/@qcompleted)/@qcompleted), 1), %insert( ".", %eval( (@qbonust*100/@qcompleted)/@qcompleted), 1)))}

Used Pretty Print
Syntax Colourizer

I wasn't sure what your variables would look like. But after testing it with small numbers and moving the location of the *100 I think this worked. For example, with 4 and 5 it returned 0.16. With 2 and 5 it returned 0.08. Hope it does what you need. Very Happy

SubAdmin Joined: 18 Nov 2001 Posts: 5182

#VAR qbonusta {%eval(@qbonust*100/@qcompleted)}
#VAR qbonusta {%insert(".",@qbonusta,%eval(%len(@qbonusta)-2))

It is much shorter this way.

Posted: Sat Sep 24, 2005 8:01 pm

At this moment, the current variable values are:

Apprentice Joined: 20 Sep 2005 Posts: 135

They're both working properly.

Vij's evaluated (49*100)/637 = 7. It then inserted the "." in front. (needs to add ".0" if length < 2)

Mine first evaluated 49/637 = 0.07, which evals to 0.
It secondly evaluated (49*100/637)/637) = 0.01, which also evals to 0.
It then evaluated 0 is < 2 digits in length, added ".0", and concated them to 0.00.

I missed that you were going for the modulus. I had written them as divisions. But since I moved the *100 they return the same results. Are the formulae you are using correct?

SubAdmin Joined: 18 Nov 2001 Posts: 5182

You original use of the modulus operator suggested to me that you would be seeing values greater then 1. It is quite simple to deal with this.

#VAR qbonusta {%concat("00",%eval(@qbonust*100/@qcompleted))}
#VAR qbonusta {%insert(".",@qbonusta,%eval(%len(@qbonusta)-2))}

If you are going to be tracking data for long periods of time it is possible for this formula to fail as @qbonust approaches 1% of the integer limit roughly 240 million. From the values you have supplied I do not see that as being very likely.

Posted: Mon Sep 26, 2005 10:49 am

Hrmm not actually sure what modulus means :P Here's an examples of where I may use that formula:

In a quest tracking script, to track qps/quest (such as in the above). I use it to divide qps earned by total quests. The number could be 100 quests and 1500 qps, and I'd want the output as "15" or "15.00" But if it was 100 quests and 1526 qps, I'd want "15.26" -- my formula works for both, but in a third example, it could be 100 quests for 1506 qps .. instead of displaying 15.06, it displays 15.6. Because the formula I am using does 2 math equations -- one for the integer, one for the decimal, it returns the second result as 06, then drops the 0, in the above example.

The problem with the second formula you gave mem vijilante, if that it always returns at 0.(X). So if it's 29 quests and 2 qps, it returns correctly 0.06 (should be 0.07, as the actual number is 0.689xxxx, but close enough). But now if it's 29 quests, 50qps, it returns 0.172 instead of 1.72 (because of the %concat). I'm looking for an equation that will always return the actual value out to 2-3 decimal spots.

My apologies if I didn't make this clear in my original post. Slaem's first suggestion appears to work, once I make the numbers larger (as he said above). In the majority of the scripts I use this formula in, the result returned will have a number between 3.7 and 1000 (various scripts). The only exception to this (but the one that matters the most, because the number is so small) is the stuf I posted in the original post. That evaluates "lucky bonus" qps, and generally will involve very small numbers.

Is there a simpler way to do this, just using the %float command? I have tried (amongst other things) using %float in this several times, and I cannot seem to get it to work right. However, I know I used %float once in the past (a long time ago) to get a similar result. Unfortunately, I no longer

I went back and re-tried everything you guys have posted. Slaem's original string, when working with large numbers, is pretty much what I was looking for, but on smaller numebrs the return of 0.00 is not :/

Thanks a lot for the help with this :)

Posted: Mon Sep 26, 2005 3:02 pm

See if this works

#ALIAS lquest {
#MATH qptemp ((@totalqp*100)/@totalquests)
#MATH averageqp (@qptemp/100)
qpfloat=%mod( @qptemp, 100)
#IF (@qpfloat < 10) {qpfloat=%concat( "0", @qpfloat)}
%1 ~@Rquest:~@G @questmob ~(@questarea)~@M time: @questlength~@C reward: @questreward@tierbonus@doubleqp@questbonus~+2~=%eval( @questreward@tierbonus@doubleqp@questbonus+2)qp@questpracs@questtrains@questtp~@Y average: @{averageqp}.@{qpfloat}qp/quest over @totalquests quests.~@w
}

SubAdmin Joined: 18 Nov 2001 Posts: 5182

I remember one set of formulas I used that ran a series of #IF checks on the size of numbers it was working with and then used a formula most suited to give the best precision. I don't personally support the use of %float since it often returns numbers with unacceptable inaccuricies. What I have written is a quick and dirty formula that assumes a specific range in the numbers, in both cases. In the latter case this should actually be:
#VAR qbonusta {%concat("00",%eval(@qbonust*100/@qcompleted))}
#VAR qbonusta {%insert(".",@qbonusta,%eval(%len(@qbonusta)-1))}
This is because %len returns the total length ("12345" returns 5) then %insert puts the character at the specified point ans pushe the rest back (%insert(".","12345",4) returns "123.45". So there was an error in my formula.

Apprentice Joined: 20 Sep 2005 Posts: 135

@Atreides_096, did you mean to divide by @qcompleted twice? That would explain the very small results evaluating to zero.

#VAR qbonusta {%concat( %eval( @qbonust/@qcompleted), %if( %len( %eval( (@qbonust*100/@qcompleted)))<2, %insert( ".0", %eval( (@qbonust*100/@qcompleted)), 1), %insert( ".", %eval( (@qbonust*100/@qcompleted)), 1)))}

Check out %round() for more accurate results.

Used Pretty Print
Syntax Colourizer