BookmarkSubscribeRSS Feed
genemroz
Quartz | Level 8

Last week I submitted a question on ranking of scores and received a very helpful response from s_lassen. I subsequently learned that the ranking rules had changed and I am now back with a request for additional help.

 

The attached code produces output that, I hope, illustrates the issues. The column labeled WantedRank is the desired outcome of the Ranking Code which is labeled NewRank. The Agreement column indicates agreement between the two. The Explanation column is the rule that governs the ranking.

 

Players are divided into groups based on the previous ranking (OldRank). There are 4 active players (Registration=1) in each group but there can be several Inactive players (Registration=0 and Score=0) in any group.

 

The ranking rule for the 4th place player when there are Inactive players in the group is the stumbling block. The rule is that the 4th place player in the group must move down in rank below any Inactive player in the group AND below the 1st place player in the lower group. I haven't figured out how to capture this rule in the code.

Thanks in advance for taking the time and effort to consider this problem.  Let me know if you have any questions.

 

Regards,

Gene

 


data work.RANKINGDATA;
  infile datalines dsd truncover;
  input OldRank:BEST12. Player:$26. Registration:BEST12. WantedRank:BEST12. Scores:BEST12. Groups:BEST12.;
  format OldRank BEST12. Registration BEST12. WantedRank BEST12. Scores BEST12. Groups BEST12.;
datalines4;
1,Jo Comstock,1,1,42,1
2,Andrea Dilger,1,2,39,1
3,Sonja Drinkwalter,1,3,38,1
4,Lynn Manns,1,5,29,1
5,Linda van Son,1,6,38,2
6,Nancy Popenhagen,1,7,36,2
7,Ann Elliott,0,8,0,2
8,Jan Raugust,0,9,0,2
9,Sara Meline,1,4,40,2
10,Renee Salko,0,10,0,2
11,Karen Weldon,0,11,0,2
12,Georgia Graeff,1,13,30,2
13,Patti Holden,1,12,40,3
14,Cris Silkman,1,15,35,3
15,Jane Wood,1,14,36,3
16,Jennifer Roberts,0,16,0,3
17,Pat Chernow,1,18,21,3
18,Roberta Diles [sm],1,17,44,4
19,Norma Lozano,0,19,0,4
20,Tammy Dana-Bashian,0,20,0,4
21,Lori McLinton,1,21,43,4
22,Beth Weaver,0,22,0,4
23,Katie Claypool,0,23,0,4
24,Cathy Dosch,1,24,38,4
25,Michelle Dileonardo,1,26,36,4
;;;;

/*Code for Ranking Ladder Hierarchy Pickleball Results*/
/*Developed from recommended approach by s_lassen on the SAS Communities User Group*/

/*Create Rank of Scores by Group*/
proc rank data=rankingdata out=rank1 descending ties=low;
by groups;
var scores;
ranks Result;
run;

/*If player did not play then is Inactive and Result set to 9*/
data work.rank2;
set work.rank1;
if registration=0 and scores=0 then result=9;
run;

proc sort data=work.rank2;
by groups result;
run;

data regroup;
  set rank2;
  Length Explanation $150;
  /* winner of lower group moves up to above 4th place in higher group */
  /* move winner */
  if result=1 and groups>1 then do;
    group2=groups-1;
    result2=4;
    end;
  else do;
    /* make room for winner */
    result2=result+(result>3);
    group2=groups;
    end; 
    
/* Explanations */
if Groups=1 and Result=1 then
explanation='Winner of Group 1: New Rank=1';

If Groups NE 1 and Result=1 then
explanation='1st Place in Group: Ranks ahead of 4th Place higher Group';

If Result=2 then
explanation='2nd Place in Group: Ranks ahead of all players within Group';

if Result=3 then
explanation='3rd Place in Group: Ranks below 2nd Place in Group';

If Result=4 then
explanation='4th Place in Group: Ranks below winner of next lower Group and below any Inactive Players within Group';

if Result=9 then
explanation='Inactive: Can only maintain rank or go down in rank if a lower ranked player moves above them by another rule';
run;

proc sort data=regroup;
  by group2 result2;
run;

data want;
  set regroup;
  Length Agreement $3;
  NewRank=_N_;
  drop group2 result2;
  
/*Differences in Rank*/
if  WantedRank=NewRank then Agreement='Yes';
else Agreement='No';
run;

proc sort data=want;
by oldrank;
proc print data=want;
Title "Comparison of Manual Ranking with Ranking Code";
var Groups Player OldRank Registration Scores Result NewRank WantedRank Agreement Explanation;
run;
6 REPLIES 6
ballardw
Super User

Can you show what the expected result for process should be given that data set?

genemroz
Quartz | Level 8

If you run the code, the expected result is in the column entitled WantedResult.  

Thanks for taking a look at this.

 

Gene

ballardw
Super User

@genemroz wrote:

If you run the code, the expected result is in the column entitled WantedResult.  

Thanks for taking a look at this.

 

Gene


That implies that you code is the solution. So why are you asking a question?

genemroz
Quartz | Level 8

The WantedResult column was derived by following the ranking rules and manually entering the ranking as a variable in the input dataset.  It is not a computed result.  I did this so you could see at a glance where the computed result disagreed with the desired result.  My hope is to arrive at code that can do this ranking accurately and faster than the manual approach for so that it can be scaled to situations involving many more players.  Thanks again for taking the time to consider this question.

Regards,

Gene

Patrick
Opal | Level 21

@ballardw 

The desired result is in the "data have" wanted_result column and though not computed. 

The related question the OP mentions is here.

@genemroz 

It would help if you could explain the required logic in detail or if it uses some "standard" rules for seeding provide a link to a side that explains the logic in detail. It's at least for me too much effort to try and "guess" the logic from what you provide prior to actually creating the code.

genemroz
Quartz | Level 8

Yes, I know the rules are complex and subtle.  I know of no link that provides a better description.  Here are the rules I am working from:

1.  If you outscore a higher ranked player in a Ladder Match (Group), you move ahead of that player in rank.

2.  if you finish first in your group, you move ahead of the player finishing last in the group just above your group.

3.  No inactive player can move above an active player.  When active players lose and go down in rank, inactive players below them get pushed down as well.  ACTIVE players are assurred that no INACTIVE player can move from below them to above them in rank regardless of what they do.  At the same time, INACTIVE players must realize there is nowhere to go but down.  (Note that INACTIVE players identified in the dataset with Registration value of 0 and a Score of 0).  Note in Group 2 that the WantedResult shows Georgia Graeff, who came in 4th in the contest, stays below the 4 Inactive players in that Group.  The WantedRank of Inactive players can only stay unchanged from OldRank or go down (ie take on a higher value).

 

Hope this helps.  Thanks for trying to understand it. I know it's a bit of challenge.  I'm happy to answer any questions.

 

Regards,

Gene

 

 

SAS Innovate 2025: Register Now

Registration is now open for SAS Innovate 2025 , our biggest and most exciting global event of the year! Join us in Orlando, FL, May 6-9.
Sign up by Dec. 31 to get the 2024 rate of just $495.
Register now!

How to Concatenate Values

Learn how use the CAT functions in SAS to join values from multiple variables into a single value.

Find more tutorials on the SAS Users YouTube channel.

SAS Training: Just a Click Away

 Ready to level-up your skills? Choose your own adventure.

Browse our catalog!

Discussion stats
  • 6 replies
  • 794 views
  • 0 likes
  • 3 in conversation