Adds a 12 player cap to the distance calculation

My understanding is that Kart v1 was originally tested and designed around 8-12 players being optimal. In my time playing v1, as well as talking to various long time players people generally agree that in those 14-16 player lobbies item rolls tend to be undertuned and not providing enough catch up power or pack compression.

This MR puts a cap on the multiplier that is applied to distance calculations used for odds bracket determination. With a full 16 player lobby, the multiplier normally goes down to 0.714. With this change, the multiplier is capped at 0.857. At most a ~15% increase moving down the distance brackets in the largest of lobbies. Consider it as a 15% loose target pack compression at most, that only affects the largest lobbies.

Part of the frontrunning problem (not all of it) in these large lobbies can be attributed to 2nd rolling the more close combat items in bracket 2, when they might have needed some of the catch up options from bracket 3 for example. This will provide a subtle increase in ability to attack first for these 13 player and up lobbies. It will also subtly compress the pack of players in these large lobbies, increasing player interactivity. This will also improve player experience on small 5 lapper maps in large lobbies, which are considered poor due to the low distances involved.

I do believe this change is subtle enough to not upset Kart's general design. It's not as effective as say, the Juicebox addon. Many less experience players may not notice. We have tested changes similar to this in 1.3 public games in my server and in NVJR servers via the xItem addon, with its reimplementation of the item rolling code in Lua. The impressions were generally positive, however this was also mixed with some small adjustments to the odds brackets themselves, so will require more netgame testing in an isolated fashion as proposed here.