A few observations and clarifications:

1) Bullets of the same weight differ in their "slipperiness" in the bore.

There are differences in jacket alloy and hardness, core alloy and hardness, differences in diameter and differences in bearing surface that all change the pressure generated with a given load.

A lot of handloaders recommend you use data from the powder company, because they make the powder, so their data must be best, right? Wrong. That data is based solely on cast versus jacketed and bullet weight and doesn't take into account the above mentioned variables in the bullet.

For the last 43 years my preference has always been to use data developed by the bullet manufacturer whenever it is available. I'll use powder company data, but it's a second or even third choice and I take it with a large grain of salt.

2) Seating depth and amount of crimp have significant influences on pressure.

In small high pressure rounds like the 9mm Luger pressure can quickly exceed a safe limit if the OAL is too short as the bullet will be seated too deep and reduce the initial volume available for the powder to ignite. That can occur due to set back as well when the round is chambered in a semi-auto pistol or carbine. Set back is a bigger issue with fast burning powders than it is with slow burning powders.

In large capacity rounds like the .357 Magnum, AOL is less important as the decrease in volume in the case if the bullet is seated slightly deeper usually isn't an issue. Seat to the crimp groove and leave it at that.

However, crimp matters for two reasons. First, if the crimp is not adequate the bullet can back out of the case under recoil. That becomes increasingly an issue with higher pressure loads in lighter weight handguns. Second, if the crimp is inadequate, the bullet can start down the cylinder before the powder fully ignites and that can create inconsistent ignition. A firm crimp will keep the bullet in place long enough to reach a point on the pressure curve where you get consistent ignition. This is much more important with slow burning powders than with fast burning powders.

3) Win 296 and H110 are the same powder.

They have been the same for decades. However, there is a lot of variation lot to lot in Win 296/H110. For example, if you start looking at manuals with .357 Mag data for both Win 296 and H110, you'll never find the same maximum charge, and in some cases you'll see a difference as high as .7 grains when used with the identical bullet and other identical components.

That's the potential lot to lot variation, and consequently you need to give up on the idea that there is some "perfect" maximum load, just from effects of lot to lot variation in the powder.

4) Published "nominal" charge weights are not a precise recipe.

I'm using a military example, but it applies to all powder, but most handloaders fail to understand this. You'll see this a lot in forum discussions where someone wants to replicate something like .30-06 M2 ball ammo. Someone will state that the load for it is "47.5 grains of IMR-4895 under a 150 gr FMJBT bullet".

The problem however is that the M2 Ball projectile varied from 152 grains down to 147 grains over the years. In addition, US Army TM 43-0001-27 (Army Ammunition Data Sheets: Small Caliber
Ammunition) on pages 5-7 describes the "Cartridge, Caliber .30, Ball, M2" as using:

"Propellant Type: IMR 4895
Weight: 50 grain"

That obviously isn't 47.5 grains of IMR-4895. However, less obviously, even though it coming from the US Army technical manual, 50 grains isn't "right" either. It's just a nominal load.

When a contractor or a government arsenal loaded M2 Ball ammo, they developed a load for a large lot of powder (2,000 to 10,000 pounds) and developed a specific charge weight for that lot of powder that might have varied from 46 to 52 grains based on the bulk grade powder used. There is no "precise" load that will magically replicate M2 Ball (or any other military load for that matter). The charge was determined based on what was required with that particular lot of powder to achieve the specified 2740 fps velocity measured at 78 feet from the muzzle.

I've seen similar discussions where a shooter would try to reverse engineer something like M118LR by pulling the bullet and weighing the powder charge, then determined that it uses "43.1 grains of RL-15, based on the weight they got on their scale and a rumor that RL-15 is what was used when M118LR was developed as it is less temperature sensitive than IMR-4064.

The problem is that when Federal makes M118LR, they don't buy canister grade RL-15. Instead they buy a large lot of a bulk powder that has the same chemical composition, structure, and similar burn rate, and then develop a specific powder charge for that lot of powder that produces the required military specified velocity while staying under the maximum specified average pressure.

Worse, you'll see the "internet expert" specify a "military match case". Back in the day 7.62x51 NATO brass was thicker than .308 Commercial brass, even though they otherwise share the same exact external dimensions. This was done to reduce the potential for a head separation that might jam an M60 with a barrel that had excessive headspace. As far as I know that's still the case with M80, but Federal uses their Gold Medal Match case in their "match" M118LR load. Interestingly enough, I bought 500 pieces of new in the package unfired Lake City stamped brass that had had the same average weight and volume of Federal GMM brass. Guess who makes that brass - and while it is stamped Lake City and Match, it is commercial brass with a lot more volume than the old Lake City Match brass.

Details matter, and you need to understand that a published "nominal load" is not a precise recipe for cloning military ammunition.

5) Powder variation only gets worse once you move away from canister grade powders.

There is less variation in the canister grade powders sold for the purpose of handloading, but there is still some lot to lot variation, which is why you are always supposed to start at least 10% below the maximum published charge weight.

Where handloaders get into test pilot territory is in using surplus powders. These are not canister grade powders and there is a lot of variation. For example .223 Rem /5.56x45 NATO handloaders will talk about BLC-2 being the civilian equivalent of WC 846 and H335 being the civilian equivalent of WC 844. What they fail to grasp is that there used to be just WC 846, and WC 844 was developed as a new specification on one end of the broader WC 846 specification. Look at how far apart BLC-2 and H335 are on a powder burn rate chart and understand that the WC846 spec treated them as the same powder.

WC 846 is still a broad specification and canister grade BLC-2 is just one point in that range. WC 844 is still much broader than H335 canister grade powder. Consequently, I cringe when I read someone pulled down a M193 ball round and proclaims you need to use X grains of WC844 with a 55 gr FMJBT to replicate M193 ball ammo.

In terms of surplus powder, there's new old stock surplus military powder, which I will use, with due regard to load development, and surplus pull down powder that I will not use under any circumstances. Pull down powder comes from rounds that were made from a number of different lots of powder along that very wide range, and you have no idea what you have and whether the mix even comes close to the original specification. Don't go there, regardless of how cheap the powder may be.

6) Use a chronograph to verify your results are in the expected ball park.

Even if you follow the recipe in the manual precisely, using the identical components specified, your results may vary a bit due to differences between the test firearm used to develop the load and your firearm. That's pretty obvious, especially when the barrel lengths are different.

What is less obvious that even with two firearms of the same make and model you can get significant velocity differences. You may find that 2 S&W 686 revolvers with the same dash number and barrel length may produce a 50 fps to even 150 fps differences in velocity with the same load.

New tooling will produce larger chamber and bore dimensions than old tooling and that makes a difference. Cylinder gap also makes a big difference in velocity. A revolver with chambers and bore cut with old, almost ready to be retired tooling and a tight cylinder gap may well be 150 fps faster than the same model and barrel length revolver made with new tooling and with a larger cylinder gap.

So...don't think you have to, or even can, match a published velocity, even if they used the same make, model and barrel length revolver as yours. Pay attention to pressure signs and stop when it looks like the pressure may be excessive, even if you are still under the published maximum velocity for that load. Personally, I back off slightly in .357 Mag when the cases start to stick on ejection, and that can vary from revolver to revolver.

Over time you'll get a feel for where your firearm performs versus published load data.

Once you have a load that works, use a chronograph, both as a sanity check to see how a lot of powder performs, and to adjust the charge weight for that lot of powder to give you the same average velocity as the load with the old lot of powder that worked well for you. You'll spend less time developing loads if you buy powder in 8 pound kegs.