/* In balanced mode, the amount of diff1 solutions per pool is monitored as a
 * rolling average per 10 minutes and if pools start getting more, it biases
 * away from them to distribute work evenly. The share count is reset to the
 * rolling average every 10 minutes to not send all work to one pool after it
 * has been disabled/out for an extended period. */

So --balance looks to the diff1 count and tries to get an even distribution.

--load-balance looks at the --quota values and distributes work in proportion to the GCD of them, counting down until remaining quotas are all 0 and the values are reset.