Group setup

What is the rationale for the group setup in the tournaments? With the current group setups there are major quality differences between the groups!

For example; 8 players; 2 groups: The 8 are ranked and grouped so that number 1,3,5 and 7 are in group A. The quality of group A is then 16 (1+3+5+7), while group  B is 20 (2+4+6+8). As the tournaments grow bigger, the differences are bigger. Currently I'm in a tourmanent with 16 groups of 6 players. Here the best group has a total ranking of 246, while the worst group has 336.

This is not a fair way to setup the groups. Instead the setup should be (8 players, 2 groups); Group A: 1,4,5,8 Group B: 2,3,6,7.
Or is there another reason for the current setup? 

No, I don't agree, the rationale behind the reason given is wrong.

Again, in the tournament I'm in, there are 16 groups of 6 players. With the current setup, the player ranked as number 16, gets a lot easier group than the player ranked as number  17. With the current setup, it is not at all ensured that the best players get to the final round, but a lot of luck.

If you're one point short of being number 16, you'll have to play agains number 1,33,49,65,81 instead of 32,48,64,80,96 - and your way to the final round is suddenly way harder.

In my setup, player 17 (who might be just as good as player 16) gets a way better chance at qualifying to the next round.

I am not sure that this is exactly what Jay had in mind.. but I was able to create an algorithm for this i am not sure if it works for groups that are other sizes than 6.. but maybe you can play around with it.. this is the player spread.. the algorithm follows... 

1 32 48 49 50 96 = 276
2 31 47 51 52 95 = 278
3 30 46 53 54 94 = 280
4 29 45 55 56 93 = 282
5 28 44 57 58 92 = 284
6 27 43 59 60 91 = 286
7 26 42 61 62 90 = 288
8 25 41 63 64 89 = 290
9 24 40 65 66 88 = 292
10 23 39 67 68 87 = 294
11 22 38 69 70 86 = 296
12 21 37 71 72 85 = 298
13 20 36 73 74 84 = 300
14 19 35 75 76 83 = 302
15 18 34 77 78 82 = 304
16 17 33 79 80 81 = 306

1.First player is assigned to first group

2. Keep assigning next ranking player untill all groups have one player assigned

3.  Reverse assignment direction and assign next player from ranking from lowest to highest

3a. Find midpoint rank from total ranked players.

3b. Keep assigning players until you get to the midpoint(should end up at highest ranked group unless i messed up somewhere).

 4. Place mid point and midpoint + 1 in the same group and then go to next group down place next two ranked players until you reach the last group

5. Place next three ranked players in the last group

6. Reverse back upwards and keep placing ranked players until you reach the last player.

This should produce a tighter knit group spread. My test of this with 16 groups of 6 with a total of 96 players with 48 midpoint gave a spread of 2 places per group max 276, min 306 .. which is much less than the previous algorithm gave..

Haha now i am doing this on very little sleep, so it may be only good in this situation but i tried it on a few different senarios and it looked fine. Let me know if there are issues, or if you have questions.

Ciao 

Luminary: "My" way is different. I sent Jay this simple algorithm for it (waiting for response):

This will give the groups:

1 32 33 64 65 96 = 291

2 31 34 63 66 95 = 291

3 30 35 62 67 94 = 291

4 29 36 61 68 93 = 291

5 28 37 60 69 92 = 291

...

16 17 48 49 80 81 = 291

Given that the number of players in the group is an even number, this will lead to all groups being of the exact same total strength.

From what I see, your suggestion will create even bigger differences in group strength if the number of players in each group is 5... 

I think, the "quality rationale" for each group should not be the sum of their members ranking numbers (1,2,3,...) but, instead, the sum of  their members ranking points (2345, 1987, 1431,...) . One rationale (not the only one!) should be, that this sum is nearly equal for all groups of a tournament round.