Quelques mots de sagesse d'un compagnon britannique, source régulière de propos équilibrés et sains :
There is no such thing as a solution that works every time in every conceivable context. There are no-win scenarios, just as there are scenarios when prevailing is easy.
Running away is by far and away the response with the highest percentage of success when applied to, say, 100 randomly created scenarios. I recall reading about a very well organised scenario conducted with simunitions (very posh and expensive paintball). Participants ranged from special forces soldiers to untrained neophytes. The problem was that of multiple armed adversaries in a structure. The most successful tactic, the one that gave the highest percentage of survival, was escape. The vast majority of those who stood their ground and fought died. The few that "won" were those who made an immediate effort to escape the intial attack and counter ambush the bad guys.
So, the top performers were a few guys with millions of pounds worth of training behind them and those who decided that fighting wasn't a good idea.
Escape is usually a "silent success" that does not get included when considering a statistical analysis of responses. Just as you can drive to work every day and get there just fine, it is not regarded as a success because the rewards for achieving it are trivial. If you are forced to make an emergency correction when driving, this is far more dramatic and becomes "life saving skill" in memory. The simple fact that sensible, everyday driving has saved your life daily every day in the last year is discounted. Good daily driving skills (awareness, being in the appropriate gear, hazard anticipation, road positioning etc) are far more useful than skid control.
Many, many, many deaths are averted each year because folks very wisely choose to escape rather than stand and fight. They don't phone up the National Antelope Impersonation Centre and say "Hey, I didn't die last night because I ran faster than the dude behind me. Chalk it up on the board." Incidences of active, successful resistance are far more likely to be recounted and remembered - making up much of the memory of responses to violence.
This tendancy to remember the dramatic also means we are very likely to remember rare incidences of failure of an otherwise reliable tactic. We risk grossly under-estimating our chances of success with Safe Tactic A, instead choosing to opt for Risky Tactic B because we remember the successes of B together with the failures of A. The memory sample is not a true reflection of facts.
To reiterate, there are circumstances where you should fight, and there are circumstances where you should flee. The dice is loaded in the favour of discretion rather than valour. Handily, this is exactly why the fight/flight reaction to sudden displays of aggression is set by default to run away - hence the shaky legs during the pre-fight phase. Millennia of natural selection shaped this tendancy - it is not by fortune alone that we choose flight over fight. The organisms that resolutely chose to stand and fight, even when escape was a viable option, failed to live long enough to produce enough progeny to become the dominant species of the time and space. They who chose to run away, lived to feed, breed and fight another day.