To know a good bounty we should consider the time frame of the bounty, the amount allocated for the bounty, the manager managing the bounty, and the validity of the project doing the bounty campaign. There are lots of bounties today that offer a reward that is almost non-existent. They are requiring months of work but the reward is miniscule. Better learn how to calculate reward base on the participant number if the bounty offer is in stake form.
All of these things would be useless from payment or allocation of coins been put up on bounty if this certain coin/project if it would be ending up on having no value.
This is why it would really be just needed for you to make out some in depth research of a project even though it wont really be giving out assurances that it would succeed but
it is really better rather than on doing nothing at all. Sticking with known managers is good since they are doing the initial part on choosing good projects specially to those
reputable and popular where success chance is high compared to others and with that alone then you could already presume out.