Tile tool to split 10 rows into two equal groups based on numeric 'Rating' value.

[GarethLiggett]

I have a list of 10 players

I tried using the Tile tool with equal sum method to split the rows into 2 groups (as close to even as possible)

This has not worked as expected. The tile tool has just categorised the rows in the order they appear, with tile 1 have a much higher rating.

I may be using the Tile application incorrectly, or my understanding of how it works is incorrect. Alternatively can I achieve 2 even groups with an alternative tool? Perhaps by listing out all possible team combinations and summarizing to get average rating across the teams.

[DataNath]

Hey @GarethLiggett, I thought this was a little odd and your configuration is spot on so had a little bit of a dig. @DavidP gave a great explanation in this thread - https://community.alteryx.com/t5/Alteryx-Designer-Discussions/How-to-properly-use-the-Tile-Tool/td-p...

Essentially, the Tile tool doesn't appear to give great results for smaller datasets, at least when in Equal Sum mode.

[GarethLiggett]

Thanks DataNath - Any other suggestions on how I could replicate this functionality for 10 rows? E.g. get a list of all possible team combinations (unsure on this process)? If I had this I could generate a MatchID and TeamID and calculate the difference between Team 1 vs Team 2 and sort by the difference ASC.

[DenisZ]

I am not so sure how simple it is using Alteryx. You could try to match all possible team combinations, but this would be a 10 factorial number of combinations. However, if you have experience with Python, then that tool can be leveraged as there appears to be multiple codes online for that. 

Some examples: 

python - Splitting list into 2 parts, as equal to sum as possible - Stack Overflow

[DataNath]

@GarethLiggett I've just quickly put together an iterative macro that splits the data in 2, sums both teams and then creates a % difference of this split. Upon outputting from the macros, the iterations are then sorted by % diff and the top one taken. At the moment, I've just set a hard limit of 50 iterations (I've ran it a few times and usually the best result in this case - 0.43% - is found within ~5 iterations anyway, though there were a few runs that went into the 20s etc). However, you could set a threshold for the % diff as a point for the macro to end if there was a result that you'd be happy with without finding the optimal. Hope this helps and please let me know if you have any questions:

(The error is just the macro hitting the iteration limit I set - don't worry!)

[Present_guy]

So..... I conclude that the tile tool will not do an equal sum grouping and eagerly await the new tool that will be added to Alteryx that will actually do this. :(