 |
[Date Prev][Date Next][Thread Prev][Thread Next][Date index][Thread index]
Re: st: RE: Re:distribution code
Hi Victor,
I have written a bit of code that I think might do the trick. I'm sure
it could be condensed and written in to a loop, but I wanted it to be as
clear as possible. I think it could be extended for more teams fairly
easily.
*---------begin code--------*
clear
set obs 7
input round
1
2
3
4
5
6
7
end
* gen initial values for 8 teams
gen A = 1
gen B = 1 + round
gen C = 2 + round
gen D = 3 + round
gen E = 4 + round
gen F = 5 + round
gen G = 6 + round
gen H = 7 + round
* fix so 9->2, 10->3, etc
replace B = B-7 if B>8
replace C = C-7 if C>8
replace D = D-7 if D>8
replace E = E-7 if E>8
replace F = F-7 if F>8
replace G = G-7 if G>8
replace H = H-7 if H>8
* for the brackets: the matchups are always AB CH DG EF
egen field1 = concat(A B), punct("_")
egen field2 = concat(C H), punct("_")
egen field3 = concat(D G), punct("_")
egen field4 = concat(E F), punct("_")
*--------end code--------*
The above gives me the following matchups:
. cl field*
field1 field2 field3 field4
1. 1_2 3_8 4_7 5_6
2. 1_3 4_2 5_8 6_7
3. 1_4 5_3 6_2 7_8
4. 1_5 6_4 7_3 8_2
5. 1_6 7_5 8_4 2_3
6. 1_7 8_6 2_5 3_4
7. 1_8 2_7 3_6 4_5
Does anyone see anything wrong with this approach?
-Elizabeth
Nick Cox wrote:
Victor has posted three versions of this problem, without
a flicker of public response that I can recall.
The interpretation that I find easiest of his problem is
Eight teams 1-8.
Each team plays every other team just once.
Evidently there are 8 * 7 / 2 matches and 7 rounds of matches. I notice
no distinction between "home" and "away".
I had a few goes at doing this in Stata, but in each case only managed
code for an easier half of the problem. The best data structure for this
is to me far from obvious. Victor wants each match to be an observation,
but I wouldn't be at all surprised if you were better off with each
round as an observation. Thus you are working with various shuffles of
"12345678" divided into pairs, with the understanding that "12" and "21"
are not distinct.
In case it is also of interest, I offer here a solution obtained "by
hand".
12, for example, means "1 and 2 play".
12 34 56 78
13 24 57 68
14 25 38 67
15 26 37 48
16 23 47 58
17 28 36 45
18 27 35 46
This took much less time than my various failed programming efforts.
Of course, if the real problem has say 16 or 64 teams, and this one is
just a toy, then this is no use whatsoever.
Nick
[email protected]
-----Original Message-----
From: [email protected]
[mailto:[email protected]] On Behalf Of Victor M.
Zammit
Sent: 14 January 2008 22:01
To: [email protected]
Subject: st: Re:distribution code
Dear Listers,
I have the following dataset
<snip>
and I need to find the code capable of distributing it in seven groups
,of
four observations (pairs of numbers) each ,such that each group would
have
eight numbers and no group may contain any repeating number
and no pair of numbers may be repetitive within the seven groups.For
example
:
1 2,3 4,5 6,7 8
1 3,2 6,3 7,4 8
.........................
.........................
.........................
........................
1 8,2 7,3 6,4 5
*
* For searches and help try:
* http://www.stata.com/support/faqs/res/findit.html
* http://www.stata.com/support/statalist/faq
* http://www.ats.ucla.edu/stat/stata/