Thanks for the A2A, and good question. IMHO…

A good roadmap:

• Alludes to the larger strategy, perhaps having several “pillars” or themes that the product must achieve excellence to achieve that overall strategic vision. If you had the roadmap for the original iPad it might have shown incremental improvements in each phase toward a vision of a mobile phone.
• It's concrete in the near term, and loses precision in the long term. The next quarter's (cycle's) product investments will be highly specific and well defined for engineering and stakeholders. Items 3 or 4 quarters (or years if your business can project that far) from now are roughly understood, acknowledging that data from your near term bets and your customers will shape the future roadmap.
• Priorities are clear, with a bias towards impact. Items that have the highest potential to pursue the objectives of your business (revenue, engagement, customer satisfaction, etc) are what everyone is intending to work on first, and items lower in your “stack” are ranked lower because you're at peace with only accomplishing a few of the top items in the worse case.
• Buy in and goal alignment is strong across your team/stakeholders/board. You've socialized and even constructed the roadmap and it's prioirities with the folks encharged with implementation. Engineers and designers believe in what they are building and even advocate for it within the team. Stakeholders believe in your pitch for the future and generally understand why you're doing what you're doing next.
• You've secured and marshalled the resources to make the roadmap achievable. It's not a wishlist that can't get done. There's a rough sense of the effort for completing your near term objections in a timeline that's reasonable to achieve impact. The people implementing it generally agree or have participated in estimating the time required. You've secured headcount / funding to hire more people along the way to accelerate your plan if it calls for that.
• There are no sacred cows, but you are committed to focus. Upon getting to your next milestone, you are very open to new data and information suggesting you need to evolve (or reset) your assumptions, and the roadmap may change accordingly. Experiments you've planned are designed to prove key hypotheses and remove ambiguity with the lowest reasonable effort. That said, you are skeptical of arbitrary new ideas, and appropriately resist churn to land your current roadmap. If you have prioritized the most impactful items, you are a protector of your roadmap to achieve that impact, and methodically pursue changes that avoid creating distractions.

A bad roadmap:

• Is a list of features or a wishlist aggregated from various sources advocating for their needs.
• Hasn't optimized for impact adequately, and the product never seems to grow or produce results.
• Is planned a year in advance and blindly followed. (If it's hardware, this time horizon might be a little longer.)
• Changes frequently throughout the time period you're operating under (quarters, months, sprints, etc), and is subject to distractions.
• Is dictated to the individuals in charge of implementing it without accounting for their input. (Sometimes with the command to achieve in a certain deadline, or without a timeline discussed at all)
• Are ignored by the team implementing them.

I'll ponder this more, but my hope is the answer demonstrates that roadmaps are more than a list. They are an artifact that captures all the other activities a product manager is responsible for leading their product line, and represent the outcome of an incredible amount of insight, analysis and collaboration.