underestimating risk but over or underestimate a event impact on a stock?
Tough…I would say underestimate as event impact is a risk and you will underestimate risks and therefore underestime event impacts.
it depends you can be overconfident that a netflix is overpriced or underpriced
^^^ indeed…if you are overconfident that a signifigant negative event will occur and smash the stock price, I would say you overestimate event risk. If you are all bulled up on a name you will underestimte it.
Negative suprises Anchoring leads to positive suprises
excess trading and underperformance
Overconfidence leads to narrow confidence intervals, and more ‘surprises’ than the confidence interval would suggest. Excessive trading and underperformance are both common (as mentioned above), as traders think they are at an informational advantage, when frequently they are at a disadvantage. I don’t know how event risk really plays into it past the narrow CI’s though- anchoring is the trait they usually look for regarding that. You’ll underestimate the effect of the new information.
a short career,
would say OVERestimate (just my rough guess) I know some overconfident guys from the business and that really helps to have those poor performers in mind to the exam, also for Type I error 
they short (!) the index one week after fukushima and wonder why they hit the rebounce in essence (practice not CFAI) they have no idea whatsoever is going on and use a lot of ceteris paribus excuses, too, i.e., they say their trade was right but this one event did change everything. Anoterh example: a trade from me (would have deliverd 50% plus in some 2-3 weeks) did not go through because mr. overconfident knew of one larger bid side (so he overestimated this “event”). So they have not the b…ls to stem the vola until the profit materializes which supports the narrow confidence intervals enough philosophized…