the last person up who's skirt you want to blow smoke is your own. I would not give all 3 scenarios. Pick the one you have the most confidence in (after all, you are going to be held accountable to hit or beat those numbers) and present that. You can certainly mention that you've done a lot of modeling and you understand the range of possibilities. Potential investors are going to do their own modeling and they may want to explore the range of possibilities in which case you are prepared.
The planning is more important than the plan. You have to know the numbers inside and out, upside down and backwards (for your own purposes) and you need to have faith in your assumptions. It's not that the assumptions won't change, but you need to have confidence that you've done your homework thoroughly and that you understand all the levers and which of those levers/inputs are most sensitive (have the most dramatic effect on results). you are putting your own time, money and reputation on the line here too.
As you mention, revenues are hard to forecast. My guess is your selling model is b2b. I would take the scenario that i have the most control over and work the numbers from there. Understanding the size of the addressable market is only a small part of the equation, albeit an important part. If the market is there, then the most important part is your customer acquisition plan and associated costs and growth assumptions. Things like virality and conversion are very hard to accurately forecast. I like to use a 'feet on the street' selling model as my 'worst case' customer acquisition model and go from there. Things like number of sales reps, time and cost to recruit and train, ramp time, territory size (how much territory they can handle and cost to access), deal size, etc. are easier to quantify and should present you most costly scenario ('worst case'). If you can make the revenue numbers work using a 'feet on the street' model then you can have more confidence in your revenue numbers and then model how you later make this revenue model more efficient (add a distribution partners, add inside sales, add self-service model, etc.). You have things to learn and the best way to learn them is face to face with prospects. This may not be as 'sexy' as 'growth hacking' or 'viral loops', but it's a model that is far more predictable and controllable in the early days and it's an excellent source of input for when you start to automate your selling machinery.