Ok,

Let's assume horses with 59kgs or more win 3% of races, but have a strike rate of 1/8.

Sample

1000 races with 12 starters on avg.
12,000 starters in total
3% of 1000 is 30 winners
Therefore 240 starters with 59kgs using a 1/8 strike rate.
240/12,000 = 2% of all starters have 59kgs+

Either way the % of races won is not the real issue, its the strike rate of heavily weighted horses when they start.

You probably need 2-3 years of data given that not many horses start with 59kgs all that often.

Avg dividend needed would be $8+.