Ultimately, I use a combination of both.
First I will try to identify a safe range of available charge weights for the powder and bullet I have chosen by consulting at least 3 reputable sources. I will average the minimum and maximum safe load, and use that as a baseline. Using the maximum safe load as a 100% value, I will then load a single round at the 90, 92, 94, 96, 98, 100, 102, 104 and 106% charge weights. 106% may seem high, but I have routinely been able to go north of 110% of published maximums without detecting pressure signs at normal conditions, but I have rarely found a usable accuracy node past 105%.
I will then use these single shots of each charge weight to determine what my ACTUAL maximum pressure is. I will fire them in a single ladder test pattern at 300m, max range I regularly have available. I will note any accuracy nodes that this test reveals, but I have found only a weak correlation between accuracy nodes using the ladder test vs ocw, but I find the OCW results translate to far better accuracy at long range shooting.
I will then recalibrate my minimum and maximum safe load based on actual data from the range, and conduct essentially an OCW test, except I will load 4 shots of each charge weight at 95, 95.7, 96.4, 97.1, 97.8, 98.5, 99.2, 99.9% of my maximum charge. I find 4 shot groups are a much more reliable sample size than 3, which gives me an added buffer if I have a total brain fart on the trigger. Just can't know much about a 2 shot grouping... I would much prefer to fire 10 shot groups, but I would end up burning through all my bullets doing load development and never get to actually compete.
Because I have a positive indication of my maximum charge, as opposed to the mathematical guess, I will not load a charge that exceeds maximum as recommended by OCW.
Personally, I think the most valuable aspect of the OCW test is the method of shooting the test groups, ie shooting one from each charge weight moving to a different target after every shot, and then repeating for the 2nd shot from each group etc.
We probably all typically make the lowest charge weight batch first, then the next, and put them in the bullet box in nice tiny rows from left to right, and like to shoot them the way we made them. Warming up for the day, settling down, and then getting tired, day gets hotter and brighter by the end, so many variables changing constantly. I always used to wonder if the medium fast groups were better because they were in a sweet spot, or if it was because they simply weren't my first or last shots of the day.
Since discovering OCW, I have tried several different ways of randomizing the sequence without loosing track of what bullet is what, and I find there is a lot less noise in fired groups, and the results are much cleaner and easier to analyze and interpret when you just don't know which group of bullets relates to which charge weight. I would have fallen in love with OCW just for the more reliable method of firing the groups, let alone the math on choosing the right charge weight.
My current method of prepping for a test involves assigning a letter starting with the letter K to each charge weight, drawn from a hat and write down which charge weight correlates to each letter. I will then load my 4 shots of K, 4 shots of L, etc. I skip the first 10 letters of the alphabet to avoid any arbitrary psychological attachment I might have with letters A through J. (side note, I have seen friends use a similar method, but with 10 different loads for 3 different rifles, and group C was always the tightest group, 10 out of 10. It also happened to be the first letter of his name. More often than not this best group would not have been predicted to be the best, and confused his data set.)
When I go to the range, I will deliberately leave behind my notes as to which charge weight correlates to which letter, and just shoot the bullets left to right in the normal way. When I get home I will then measure the groupings for both distance from POI as per OCW method, AND vertical dispersion, and then match the groups up with their actual charge weight and sort them according to speed. 90% of the time the group in the middle of the OCW is also the one with the tightest vertical, and those few times that it isn't it is just over or just under the OCW. In that case, I will split the difference and move my OCW half a tick towards the group with the tightest vertical.
If the group size from the OCW group is under 1/2 MOA (it usually is, best so far is .33MOA), then I am done, and no other refining of the load is necessary. I am a 3/4 MOA shooter on my best day, and would much rather spend the extra 50 rounds actually shooting than developing a load that might save me a point or two in competition.
At the end of the day, both Ladder and OCW have their strengths, and I think the ultimate answer to what is better is somewhere in the middle. Whether you prefer one or the other, a personal mad scientist mix of both, I am absolutely certain that you need to blind yourself as much as possible the differences between the loads, so as to not be able to build any expectation towards results, which can influence how much you concentrate firing shot to shot, group to group. And also, you need to fire the groups sequentially, so you are spreading the uncontrollable variables of the day, wind, light, fatigue, etc, as evenly across all the groups as possible.
) My son tried his load last week in preparation for this fall hunting, and accuracy is exactly as before in a + 20 c increase in ambient temp. I realize that may not be amazing for most, but for myself this indicates that the process outlined works well and provides repeatable results and in a manner quicker than I was able to do it before. *** edit That says something, I haven't been doing this that long. ***
