Perhaps, in addition to explaining your idea, please tell me where my approach is wrong:-

I think that for any given year a conservative approach should give a result different from an aggressive one. Therefore you need a table with years, say, on the X axis and Con/Real/Agg on the Y axis. If you would have such a table you could place the 20% benefit in the first year in relation to the Condition (risk-taking preference).

In short, if you wish to let your graphic display the effect of both Condition and Year it would be your table's duty to establish a relationshop between the two. This isn't the case, and therefore you have difficulty. To resolve the problem establish the relationship.