Investors now see other investors’ gains on trading apps, on social media, and in the testimonials of financial influencers. We study how such information affects investment decisions and whether it matters if the gain is attributed to luck or to skill. In a laboratory experiment with 739 postgraduate management students in India, 81 percent of whom already invest, participants allocate an endowment across cash and four risky assets in three incentivized rounds. Treated participants learn that a real participant from an earlier session earned about forty percent more than the endowment, and the experimenters attribute this gain to luck or to skill; a third arm adds a goal-selection prompt to the luck attribution. Peer success does not increase risk-taking. When the gain is attributed to luck, participants move about six percentage points of the endowment out of the most speculative asset and become more loss averse; attributing it to skill has no effect, and the goal prompt cancels the luck response. Participants are also strongly overoptimistic about their own portfolios. They state a 51 percent chance of a forty percent gain whose objective probability is 3 percent, and they pay more for a claim on their own portfolio than for an objectively better lottery. Peer success raises this optimism further, at a cost to their earnings. Labeling displayed gains as luck is a cheap way to temper speculation, while inflated beliefs about one’s own prospects call for disclosures of a different kind.