I think Trinz’s idea is a good one, and here is my reasoning:
In the higher levels of this game there are 3 acceptable levels of game shape – 9 which we always try our best to get to, 8 – which is most common, good but usually not good enough for the more difficult games, and 7 – which means – “oh my god, my team’s game shape really sucks”. Game shape of 5-6 is so bad that most of the serious users in the game will do anything to avoid them, even losing a game. On the other hand you can very easily get to game shape 7 – one week of over play (not 48*3, but 30-35 *3 is more than enough), or even just a couple of weeks of bad luck with the random factor.
The suggestion to open the game shape to a 1-20 scale bar, can (if deployed wisely) diminish this issue – because now the affects will be more moderate – if in the current system you can have a decline of one or two levels (which equals 2 or 4 levels of decline in the proposed system), now you can have decline of 1 or 2 or 3 or 4 levels at once, obviously the probability of 4 will be less than 2 in the old system, thus making room for more subtle changes – today most users will give a game just not to screw with the game shape –leading to a lot of uncompetitive games in the last third of the season – some can affect critical issues such as the identity if the relegated team and so on. In the new system this dilemma – game shape or 3 games in a week will be less obvious, leading to a more competitive league.
Bottom line: game shape is important thing to keep in high levels, but today it has to much weight on tactical decisions. this offer may change that.