Recently a friend of mine was trying to decide between a 36" and 38" pilot chute for his next rig. Fortunately, for this sort of thing there are all kinds of nifty charts to tell us what to do, right? Advice on forums abounds, and delay / PC size charts can be found in many places. Of course, even The Great Book of BASE features a full page guide to PC selection.
But, in reality, there are so many variables that charts and advice can only provide a very rough guide - not a precise answer. And in fact, what you will realize as your experience increases, is that, if you jump a variety of equipment, your PC selection for the same jumps will change over time as you discover what works best for you.
That is to say, you will decide what PC works best for you for each delay and gear configuration by actually using it, not by reading a PC chart.
Let's say that you're looking for a PC that you plan to use for your next trip to a sub-terminal slider-up cliff site. You have a tracking suit, a new rig, and you have limited experience with slider up objects. Most of the charts will tell you that a 38" PC is recommended for delays in the 6-8 second range, which is what you'll probably be doing. And, most of the time, for most jumpers, a 38" PC is going to work well.
A chart does not provide definite answers for every delay. This is because there are no definite answers, due to the nearly limitless variables that exist in jumper configuration, conditions, object height, and equipment.
To give just one example, an all F-111 PC in a 36" size may actually exert less pull force than a ZP PC in a 32" size. Therefore, charts can only be a rough guide, and they normally assume the use of ZP or at least majority ZP pilot chutes.
Take into consideration the size of canopy that you are jumping, and your rate of acceleration which is influenced by your body weight and clothing. If you are jumping an ultra-light canopy in a smaller size, then you may not need the "standard" 38" PC which the locals deem mandatory for their local cliff from which they usually take a 7 second delay. In fact, you might experience better performance from a 34" ZP pilot chute than you will with a 36"-38". If you are jumping a larger, heavier canopy, then you may definitely need to use the standard 38" PC. Be aware of your equipment configuration and choose your PC with the knowledge that the perfect size for your jump may not be the same as everyone else's.
Furthermore, as discussed in the book, prepare your equipment for your intended delay, not only for the height of the object. If you decide to take a 4 second delay from an object at which 7 seconds is possible, then pack for it, and don't change your plan.
Pilot chute size is critical. A too large PC will cause excessive center cell strip, which will deform the packjob, perhaps asymmetrically, and thus (potentially) negatively affect the opening and heading performance. A too small PC will increase the timeline of the deployment process, leading to a critical loss of altitude, and will also increase the amount of time that the packjob has to rotate before beginning line stretch.
In general, BASE pilot chute sizes are commonly larger than necessary, for reasons of caution. It is commonly thought that it's much safer to have a pilot chute which is a little too big than one which is a little too small, because it's better to have an opening which occurs a little too fast than one which occurs a little too slowly. However, it is important to realize that using an oversized PC can lead to disastrous consequences, including equipment failure. In many cases where a 42" PC will be recommended, a 38" PC will function well. This is not to say that you can choose your PC size indiscriminately because 'anything works', clearly this is not the case. This is only to illustrate that you should not make the mistake of thinking that because a 38" PC works at 6 seconds, then a 42" PC will work better - it won't. Packjob deformation and too-rapid deployments are not good for your equipment, for heading performance, or for your neck and back.