Evaluating regression models help

I'm new to regression and am needing some assistance in evaluating a model. The relationship I'm trying to obtain is the relationship between credit scores, and the amount of car accidents they cause( in dollars). So for example, as credit score goes up, they should cause less accidents, and therefore less dollars paid by the insurance company.
I have a simple model that is not correct, but im using this for learning purposes.

The attached screen shot is my output. I ended up transforming my dependent variable by taking the log.

Overall, this is a poor fit. F statistics is .19. T statistic is < .0001, and T statistic for credit score is .1980.

Residual vs predicted still looks clumped together toward the bottom, Rstudent does not look 'random.'

Between the bad residual plots and R values, even after I transformed, how would I go about troubleshooting this? Since P, F, and R is so high, should I look at adding another variable?

Overall, I'm just looking for a good thought process when evaluating models. I'm starting to understand the requirements for a good model, but I don't know what to do when The model doesn't meet the criteria...even after transformation