I'd respectfully like to offer an alternative opinion to those who would advocate for releasing the thumb safety at any point in the presentation prior to the pistol's bore coming level with the intended target.
Sweeping off the safety early in the draw was a technique taught back in the 70's by some of the early luminaries of practical shooting. I believe, from having known and shot with many of these folks, that their reasoning was that the trigger finger was the only safety that matters, and that the finger doesn't enter the trigger guard and apply pressure to the trigger unless and until the conscious decision to engage and fire has been made. That sounds reasonable.
Long experience has exposed some problems with this concept and it's application stemming from the way people actually do things under stress. Watching many highly trained people, including myself and others from the old "Modern Technique" school, have ND's and even injure themselves whilst performing a speed rock or time-pressure draw got me and others of my era thinking about this subject, and we modified the doctrine regarding when it's acceptable to release the safety as a result or our experience.
The subconscious sometimes doesn't do a very perfect job of timing complex actions in a sequential manner once the stress level increases and the conscious mind takes control and / or tells it to speed the entire process up. That is a core reason why conscious intrusions can and do negate and degrade the performance of the subconsciously-directed routines we work so hard to master in training.
That is particularly true when these actions are not directly trained into a matrix that depends upon one thing being done before the next one is permissable or possible.
Let me repeat this: That is particularly true when these actions are not directly trained into a matrix that depends upon one thing being done before the next one is permissable or possible.
(As an aside, we need to work even harder to learn to keep the conscious mind out of our performance; this is exceedingly difficult for many people who's ego invest them heavily in results and appearnaces, which is why we strive to simplify everything we do to run the gun. Lots of folks can shoot like gods and perform highly complex gunhandling skills on the practice range...until they are being watched or the results really count for something.)
Jerking the trigger and anticipating recoil is a classic example of a conscious intrusion (grabbing at the trigger when the sights look right, or jerking the trigger simply to "shoot fast") that overrides our better training with negative results.
Well, we see similar failures of technique when we speed our students up past their perceived comfort zones. Ergo, a modest takeup of trigger slack on the drive towards the target can morph into contacting the trigger before the hands come together, or even firing the pistol as it is being drawn. "Not me, I'm trained too well" some will say say.
I must counter that my observations of students and competitors over the years suggests that once we give permission for the trigger to be engaged through a conscious decision to fire, it will be engaged, and earlier than we might expect or require for safety. The logical brake on that tendency is to program the a series of actions, mediated by and directed in response to appropriate decisions, that must take place before the trigger is fair game.
Now, how does this relate to the release of a manual thumb safety?
In the case of a 1911 or P-35, we can train the release of the manual safety to take place only when it's safe to do so in sequence with other actions prompted by clear decisions, and not just because we have made a conscious decision to fire right from the holster (or more correctly for many shooters, a default to fire unless we decide not to).
Nor do we want to disengage the safety every time we draw the pistol for any reason and with any intent. Indeed, firing the pistol from the draw repetitively and exclusively in practice is perhaps not the best idea for folks who carry for self-defense, to say the least. If we are programmed to fire the thing every time we draw it quickly under a little stress...you get the idea.
Which leads us to the current doctrine:
When drawing the pistol with the intent to fire from a two-handed platform, the safety MAY come off only when the hands come together, and as the final act of achieving the full firing grip as the master thumb locks down onto the top of the support thumb.
This act of disengaging the safety as we lock the firing grip at high ready does a lot for our consistency, but most important it is the key that we train our minds to recognize and follow with the presentation to the target, trigger takeup and firing of the shot upon seeing what we need to see to make that shot.
When firing from retention or some variation of the "speed rock", we train to release the safety only when the master grip has leveled the pistol to the target AND the conscious decision to fire has been made and confirmed throughout the draw. If the threat changes, so does the response, and the safety is reapplied if it was disengaged at all.
Crucial to all of this: In each case of a draw leading to a presentation and a shot, the trigger is engaged only after the conscious decision to fire has been made and confirmed, and only after the safety has been released IN RESPONSE TO THAT DECISION. The safety isn't disengaged simply upon the ordinary and bloodless decision to remove the pistol from it's holster, nor even when bringing the pistol to high ready in response to a potential but unrealized threat.
(I cite at least one study, related to me recently, by the Force Science Institute in which operators put through stressful shoot / no-shoot building search sim scenarios would frequently and repetitively touch their triggers during the excercise, despite their training and conscious recollections of not having done so.)
The sweeping off of the safety requires two conditions to obtain (whether in self-defense or on the range): First, that there is a defined and recognized threat that may prompt immediate action in defense (please see above); and second, that the potential threat is safely available to cover with the muzzle.
The trigger only becomes fair game after the safety is released, and only once a final, third condition obtains: that the conscious decision to fire has been made.
If we draw to a high ready in response to a perceived threat that does not then require lethal force, we are all safer for leaving the safety applied.
If that decision to fire is made after we come to high ready, we are at no disadvantage whatsoever if we start with the pistol on "safe" before we are required to engage to save our skin. (On or off "safe", there is no significant difference in decision-to-shot times from a high ready.)
If that decision to fire is part of a reaction compelled by clear circumstances before we draw in self-defense, of course we're in deep **** and by God we'd better be right and we'd better be sure in our gun handling actions if we are to prevail.
We all have heard the reasoned admonition not to draw unless you intend to fire, but morality and reality conspire on our behalf to make us think about this from a really honest and practical point of view: Cops and armed citizens both cover potential threats at gunpoint far more frequently than they blast away at them.
The NRA reckons citizens use firearms for righteous self-defense many, many thousands of times each year, but actual shooting fights between citizens and bad guys are so rare as to be nearly statistical non-events.
I offer this in the hope that everyone will be safer and more certain both in training and when it counts. Thank you for reading!