Forum Discussion

Rich Uncle's avatar
Rich Uncle
Copper Contributor
Jun 05, 2018
Solved

Cell Format within ABCD

Hi

 

I have some worksheet to do about the ranking of the schedule and number as reference, the steps is start from X,Y,Z. now i have done format for the X value (picture attach) for the ranking within A,B,C,D which one have highest number and should be the place at top position (number followed 1,2,3,4). Example if X value between A and D have same number, should be will be use Y value (picture attach) for reference because X value is same. Example if Y value is same between A and D and should be use Z value (picture attach) for reference because Y value are same. Now if number in X,Y, Z (picture attach) is changing to another number and of course all position will be change also, so how i want to create the format for this situation if all number will changing to another? Have any one can help me?

 

Thanks in Advance,

Regards,

Uncle Rich...

  • If I understood right, you want to rank A B C D looking at X value; if there's a tie between any X, you check Y, then Z.

     

    How many groups (A B C...) are you going to have in general? 4? Or they can be 10, 200, 5000?

     

    I created something you should check.

     

    Try change values in green cells, and check the red ones (do not overwrite red cells!). Is that the right way to check the rankings?

     

    At the end, you should hide columns F to J (so you won't see the calcs).

     

    What I did:

    First, I ranked X, Y and Z in 3 different columns (F G H).

    Second, I "merged" the 3 rankings adding rank(X) + rank(Y)/10 + rank(Z)/100; in this way I have a number like 1,14 for the Team A. 1,14 means that X value is ranked 1°, Y is 1° and Z is 4°.

    If you'll have 10 to 99 groups (letters, in your example), you have to switch that formula to: rank(X) + rank(Y)/100 + rank(Z)/10000 obtaining 1,0104.

    Changing that dividers allows the bottom ranking of 99,9999 (X 99°, Y 99°, Z 99°).

    In general, you should count the last position available and adjust the number of 0 in that formula.

    I called the I column "tot. rank" meaning the sum of these rankings (now you have to forget the points, we're talking about rankings and orders).

    Now, because 1 is the best ranking, you have to sort these "tot. rank" in a new column "Rank of Totals", but you have to sort them starting from the smallest so the new formula is rank(value;all cells;1): the last "1" means that you want to have first the smallest value (1° is 1 and it's better and smaller than 5° that is 5).

7 Replies

  • If I understood right, you want to rank A B C D looking at X value; if there's a tie between any X, you check Y, then Z.

     

    How many groups (A B C...) are you going to have in general? 4? Or they can be 10, 200, 5000?

     

    I created something you should check.

     

    Try change values in green cells, and check the red ones (do not overwrite red cells!). Is that the right way to check the rankings?

     

    At the end, you should hide columns F to J (so you won't see the calcs).

     

    What I did:

    First, I ranked X, Y and Z in 3 different columns (F G H).

    Second, I "merged" the 3 rankings adding rank(X) + rank(Y)/10 + rank(Z)/100; in this way I have a number like 1,14 for the Team A. 1,14 means that X value is ranked 1°, Y is 1° and Z is 4°.

    If you'll have 10 to 99 groups (letters, in your example), you have to switch that formula to: rank(X) + rank(Y)/100 + rank(Z)/10000 obtaining 1,0104.

    Changing that dividers allows the bottom ranking of 99,9999 (X 99°, Y 99°, Z 99°).

    In general, you should count the last position available and adjust the number of 0 in that formula.

    I called the I column "tot. rank" meaning the sum of these rankings (now you have to forget the points, we're talking about rankings and orders).

    Now, because 1 is the best ranking, you have to sort these "tot. rank" in a new column "Rank of Totals", but you have to sort them starting from the smallest so the new formula is rank(value;all cells;1): the last "1" means that you want to have first the smallest value (1° is 1 and it's better and smaller than 5° that is 5).

    • Rich Uncle's avatar
      Rich Uncle
      Copper Contributor

      If I understand correctly, you want to rate A B C D see the value of X; if there is a string between X, you check Y, then Z.
      That is correct. but I forgot to explain something, I have explained below.

       

      How many groups (A B C ...) do you have in general? 4? Or they can be 10, 200, 5000?
      Only 4 at least and 8 at most in 1 group. 3 at least but rare.


      I created something you need to check.
      Thank you and appreciated it.


      Try changing the value in the green cell, and check the red (do not change the red cell!). Is that the right way to check the position?
      Ok. Yes, that's the way to check the position but there's a little more explanations that I have not described in the exact way that is, To get the real position between A, B, C, D, first need to refer to the most X, secondly refer Y most and third refer Z at least . This is because X is the amount of points collected for the entire game of example between A and B who won the game will earn  1 point, followed by Y is the highest total winning points for each game set (example) 1 to 5 set who is first, while Z is the total of lost point. This means the total number of points Y (needs a lot) and total points Z (need to be less) to determine who is eligible for a position between 1 to 4 (if 1 group has 4).


      In the end, you should hide the columns F to J (so you will not see calc).
      Ok.

       

      I have attach the another schedule with a difference amounts and example if amount in the red pentagon is less than another should be winner. But If amount in green pentagon are same also the last refer is what i call (head to head) that is very rare due to need check current match between B and D who's win the game should be winner (example) in green pentagon between B and D, D are win the current match and the position is No 1.

       

      I hope you can help me about this. I very appreciated what you have did for me.

       

      Thanks,

      Regards,

      Uncle Rich...

      • Arul Tresoldi's avatar
        Arul Tresoldi
        Iron Contributor

        I understood.

         

        By switching the ranking order of Z values (adding ";1") you give to values the opposite valor (less means better).

         

        Then, in case of a tie, you have to check who is the direct-winner, right?

        I'm not getting how some matches are not symmetric (A vs C is 1x 2y 0z and C vs A is 0x 1y 2z; I thought it should be 0x 0y 2x aka simply switching y and z -I ignore x that is the final result win/lose-).

         

        I'm attaching a newer version with tons of calcs that I'm not able to explain in details.

         

        Column AA to AD check if there are some ties.

        Column AE to AM are a table with multiple entries that states who won against who and meanwhile count all ties between first place, second, third etcetera.

         

        Column AN to AS assign to each Team (A B C D E F G H) its appropriate opposite in the tie breaker. ATTENTION: you must name the teams with A to H, and this is not working if there are more than 2 teams in a tie (let's say A, B and F are all 1x 2y 0z). This 3+ ties results in #N/D! Error in the table (col U, "POS").

         

        Column AU to BD assigns points to a team IF it won against its opposer Team in the tie.

         

        Column BE sums these points and BF create the specular (needed to rank). Column BG add these values to the rankings summed before (before applying tie-breaks).

         

        BH is now "Rank of Totals" with all Tie Breaks applied.

         

        Hide from W to BH for a better output ;)

Resources