Short range or long range
Both. Exact ratios between stand-off and short-range weapons will depend on your overall defensive posture. Defense expenditures will be very different for Poland vs Canada (Poland who gets rolled all the time vs Canada which is exceptionally difficult to invade.) If your tech-level is restricted to throwing things a few hundred yards, that's a huge difference than projecting power at 1000 miles.
Anything that lets you hit your enemy before they can hit you is a good thing. Having the power to survive and effectively retaliate when they can hit you is a good thing too. Defense in depth is standard practice in every kind of security scenario. Depending on a single technology/barrier to keep you safe will end poorly for you.
Refine the old or find the new.
Both. Taking huge jumps in tech is very risky. Far better to refine what you have since you know that works. Things that work are low risk. Things that might work are high risk. Make lots of smaller bets on "The Next Big Thing" because you may stumble on new tech that completely changes the game, eg. firearms, airplanes, jet engines, missiles, semi-conductors, etc. Most bets will fail but that's okay since they were cheap. Further, all these smaller bets may take 30 to 40 years to be really useful. Relying on them to solve some critical problem is exceptionally dangerous. It's impossible to know beforehand what will work and what won't.
Since WW1, the United States has invested considerable sums in being able to project power to remote locations quickly. This capability has expanded sufficient to be able to hit most any target on the planet with meter accuracy in less than 24 hours. (In active war zones, the time to weapon delivery may be only a minute or two, depending...). Essentially no investment has been made into home defense such as coastal fortifications since WW2 for the simple reason that the US is situated between two large bodies of water.
Conversely, many modern European nations have only a limited ability to deploy outside their region. The deployment ratios in the first Iraq war are dominated by the US.
Difficulties in Assessment
Part of what makes this assessment so difficult is that no one broadcasts their research failures, only their successes. This considerably biases the sample to things that work. This makes it easy to think that things usually work when in reality, most experiments fail and everything is just hard.
We don't hear about all the failed trebuchets that self-destructed on first firing. Nor, do we hear about the failed British tank designs in WW1 that didn't make it to the battlefield. We only read about the evolution of successful designs that were good enough to see combat.