All the instruction for normal landings that I've seen says to fly straight in to the landing spot with a certain glide slope, bleeding off airspeed smoothly, to arrive at a hover above the spot with zero forward speed. I'll be at about 45-50kt when turning final.
Why not fly the entire approach at 65 knots and flare at the end? Specifically I'm thinking that at 65kt one could enter auto-rotation and land safely if the engine should quit?
There is a difference between the training world and the real world, even in required operator line pilot recurrent training events.
Training for initial certification is very broad but generally designed to teach a potential pilot what the basic approach should include, types of basic approaches, planning that approach and then flying the specific type of approach with appropriate control adjustments to maintain the steep, shallow, normal approach; resulting in an approach to the ground with no run-on, some run-on, running on above x-speed, etc' etc. etc.
The training scenario is controlled, with optimal conditions, some challenges, etc. etc. etc.
The training scenario will require the approach be flown to resemble some defined routine. None of that, or very little of all the conditions will apply in the real world- helicopters routinely operate in order to place, remove stuff from more or less improvised landing places. The pilot should have the skills to adapt to immediate requirements, or the judgement to seek another answer....
Which is all a long way to getting to the point of your particular question- if not "A" then it must be "B"! There isn't a one size fits all answer, every parameter on an approach is on a scale, airspeed on approach, approach angle and path, approach air speed- everything, even where the wind is versus any part of the approach, even the termination.
Fast approaches are much harder to control and involve greater risk to do... what? You're not taking fire on the approach, are you? Do you think the screaming approach protects against mechanical failure? It is taught as an answer to some mechanical issues, yes, but one is at much greater risk with big power changes inducing the failure than the risk otherwise.
The biggest risk is allowing for error in your technique- one gets too low, too fast, 'behind the bucket' the decel is a significant increase in tail rotor strikes. It's also more difficult to be precise in your skid placement while you're dealing with pitch attitude, approach angle, power and speed adjustments in the last seconds where precision is everything.
Oh, and the power thing is a very real issue- even with careful calculations of weight and power and a hover check at altitude prior to approach initiation, it's a all a guess as to how much power is available and how much you'll need to terminate as planned. A slow approach gives you time to evaluate, adjust or abort. It's real exciting (and dangerous) to discover that you don't have the power available to do whatever you planned in the low and slow part, but much easier to abort than after a high speed approach and aggressive decel.
The power issue carries into turbines (other governed engines, maybe?) if you're maneuvering aggressively in the last seconds- turbine wind-down does indeed occur in which the engine maintains rotor during the airspeed decel, but the gas producer drops below the ability to recover significant power as quickly as you need. It takes time to move from 90% to 100% NG. The NR goes from you maintaining high green to out the bottom as you pull pitch to avoid impact. Been there, did that, and fortunately had altitude around the intended point of landing to fly out of the situation, nursing pitch to get back into the green.....
Those are big issues. Add that the wind is an uncontrolled, variable and critical commodity in your approach. If it's not where you really need it to be exactly when you need it to be, you need to be able to adjust. If the wind is in the wrong place when you pull pitch after, or worse yet, during an aggressive decel, you could find yourself in the dreaded vortex ring state without time and altitude.
The best approach has no pitch attitude or power changes after you align with your termination point, or minimal changes. Perfect is NO CHANGE in pitch, power, yaw or roll- you roll out at wto, three hundred, 50-65 knots and just slide down the wire to a zero airspeed or run-on speed. Besides being the easiest, most recoverable method, it allows you time for your low recon, a chance to see what's going to be a factor at termination. I hear "Oh, I do a really good high recon first!" A pilot seldom hits an obstacle seen in the approach...
Slow and boring is how you get old.