Average calculations help (how to handle DNF, +2 times)?

[Gerra]

Hi there,
I'm making a lovely timer app for iPhone, now I want to refine my average calculations.

Let's assume the session below:

00:10.00 DNF
00:12.00
00:13.00
00:14.00
00:11.00 +2

What is the best single for session here?
10.00 with DNF, or 11.00 +2, or 12.00?

What time to exclude from rolling average calculation as best?
10.00 DNF, 11.00 +2, or 12.00?

What time to exclude from rolling average calculation as worst?
10.00 DNF or 14.00?

Hope you can get the point of the question, please help me out here, I'm not sure about the correct answer.

 

[Ranzha]

Your times would become:
DNF, 12.00, 13.00, 14.00, 13.00
DNF is the worst result, and 12.00 is the best result.
Therefore, (13.00 + 14.00 + 13.00)/3 = 13.33 average of 5.

 

[tx789]

An AVerage of 5 you take away the best and the DNF is the worst and 11+2 counts as 13 or
(10(DNF)) 12.00 13.00 (14.00) 13.00+
+2 can be said as 13+ (11.00+2)

 

[Gerra]

Thanks.

So if I simply count +2 as add 2.00 seconds, and forget about the original result.
I may treat DNF like a really big time, like 59:59.999, and do the regular calculations with these times.

Then I don't have to test for these values, just use the given second values.
Thanks again.

 

[Sa967St]

I may treat DNF like a really big time, like 59:59.999, and do the regular calculations with these times.

Well, if you have two or more DNFs in an average then you have a DNF average.


edit:

http://worldcubeassociation.org/regulations/#events

9f) The results of a round are measured as follows:
9f3) All counted results and averages/means are measured in natural numbers, with averages/means rounded to the nearest tenth (x.04 becomes x.0, x.05 becomes x.1).
9f4) DNF (Did Not Finish) is the result if the solve was disqualified or unfinished.
9f8) In 'Average of 5' rounds competitors get 5 attempts. Of these 5 attempts, the best and worst attempt are removed, with the mean of the remaining 3 attempts counting.
9f9) In 'Average of 5' rounds one DNF or DNS is counted as the worst result, with more than one DNF and/or DNS the average is DNF.

 

[Gerra]

11.00 DNF
12.00 DNF
11.00
08.00 <- Whohh a new personal best!
10.00

So you say this session average is DNF?

 

[cubernya]

11.00 DNF
12.00 DNF
11.00
08.00 <- Whohh a new personal best!
10.00

So you say this session average is DNF?

Correct.

 

[Lucas Garron]

It doesn't matter what the original time is, it matters what the result of that solve is (time + penalties). There are some disagreements, but most cubers would agree that 11.00+2 is simply 13.00 (or 13.00+ if you want to note that it was a +2, but that notation sometimes confuses people); the WCA doesn't even record pre-penalty times.

EDIT: It originally seemed that I was saying the practice of *recording whether a solve is +2* is falling out of use. That's not the case, but recording it as 13.00+ is troublesome because it confuses a lot of people (i.e. is that a 13 or a 15?). It's probably still good for timer programs to use, though. (At competitions, we write "11.00 + 2 = 13.00", anyhow.) Just look at how qqtimer does it.

And DNF simply acts like infinity. Hence more than one DNF in an avg5 is a DNF average.
(DNS is also infinity, just semantically different notation for particular solves. An average cannot be "DNS", though, unless it wasn't started at all.)

 

[Gerra]

Fine.
I'll simply define a secondsWithPenalties property for every time object, and calculate averages from them.