Some interesting comments and I certainly agree with Lomcevak and Dook - Automation has proven on multiple occasions to not be the answer. Instead training, situational awareness and most importantly understanding of the individual aircrafts performance is key.
Low level aerobatics are in my view not dangerous; they simply expose you to more risk. If that risk is managed via the above, then technically one could initiate a loop (for example) from takeoff - assuming it was sensible based on the aircraft performance available.
One must first differentiate between “competition” aerobatics and display aerobatics. In competition the expectation is a ‘perfect’ loop - a constant radius, where the entry and exit speeds are the same. Skill in flying this manoeuvre is key. A slight error could result in recovery below the start height. If the margin isn’t there you will crash and thus why competition aerobatics are not generally flown at low altitudes below 500ft. Display aerobatics are not scored. A flying display should be flown to show off the aircraft, not the pilots skill. Manoeuvres need not be textbook, simply safe and ideally pleasing to those watching. As such, a loop will probably not have a constant radius - the first half flown to achieve altitude in excess of the minimum required to complete and the second half flown to make advantage of the assistance of gravity to increase turn radius and pitch rate. On the way back down, once assured that the loop can be completed safely, the radius can be gradually increased thus allowing energy to be regained and the pull managed to arrive back at the start altitude. Not all aircraft can do this, hence it is important to know the limits of the machine. A chipmunk for example will generally lose altitude in a loop. That’s not to say it can’t be done - it can, but not ending with the same energy package that you started with. In stark contrast a typhoon can execute a loop from extremely low speed / energy - it has the thrust to gain the required altitude over the apex and the pitch authority on the way back down to successfully complete the manoeuvre. However that type of loop is not the same as a competition manoeuvre, rather than prescribing a circle in the sky it will instead be more of an ‘egg’ shape.
Ultimate performance is not however, the be all and end all. Regardless of whether doing low level aerobatics in a tiger moth or an F22, energy must be managed correctly. There is always talk of too low to complete, too slow, didn’t make gate height etc. This is usually not the full story - it is almost always just energy mismanagement. The video above is a sad example. The manoeuvre (which was not a loop) was not conducted from too low or too slow and probably had the required gate altitude to complete. It was the mismanagement of the energy state of the aircraft that resulted in the accident. I will caveat that I do not know the details - I have simply seen the film and as a fast-jet aviator, understand how modern aircraft fly.
An aircrafts wing will only fly to a certain angle of attack, exceed that and under loaded flight (ie greater than 0g) it will stall. Simple aerodynamics. A modern fighter aircraft can usually exceed 30 alpha before this occurs compared to a chipmunk which is a fraction of that value. However you cannot always just pull to max performance alpha that will give you minimum radius - if you have excess airspeed you may encounter the load limit first. In a modern fighter, the load limit is effectively the G-Limiter. If you can remove that then you as a human become the limiter (ie. maintaining consciousness) or the structural integrity of the aircraft. Whilst i believe the typhoon in the video may have had the altitude to recover from the split-s he executes, he has far too much energy in the form of airspeed. On the way down the aircraft is max performing but due to either being g-limited (I assume in the region of +9g) or physiologically limited, it cannot achieve the radius of turn required from the planned gate height.
My stab at explaining some of the characteristics of aircraft performance barely scratch the surface of what is a very complex subject. We as humans will always error. The key is to learn from others and strive not to keep repeating the mistakes made in the past.