I ran some numbers in Quickload for the 69 SMK, 26" barrel, using my one of .223 case volumes for Lapua brass (30.6 gr) and the factory preset burn rate for IMR4895:
The predicted OBT node (#4) for a 26" barrel is 1.1905 ms. In my hands, every good load I've ever worked up fell very close to an OBT node barrel time; usually just a very slight bit slower. According to the QL prediction, 24.7 gr IMR 4895 would put you very close to that (barrel time = 1.195 ms, lower right corner of image), with a velocity of just under 3000 fps. At that charge weight, the predicted pressure is good (a little bit under MAX) and it gives a 99.2 % fill ratio, meaning the case is nice and full, not too low and not highly compressed. These are exactly the conditions you want to achieve the most efficient powder burn, which is generally conducive low ES/SD.
The most important thing to note is that the parameters I used (case volume, case trim length) were numbers I copied from one of my .223 rifles for Lapua brass, and the powder burn rate was the factory preset value. The predicted OBT load could change slightly depending on how different your actual values might be. However, it shouldn't change so dramatically as to put that charge weight region of your test completely out of the ballpark.
I would go back and retest charge weight using something like 24.5 through 24.9 gr in 0.1 gr increments, if at all possible also getting some velocity data for those charge weights. The critical reason for charge weight testing using QL to predict OBT nodes is to define the charge weight window that gives you consistently low ES/SD, so that you can pick your charge weight somewhere in the middle of that window. This will give you a charge weight that is the most forgiving with regard to external factors such as temperature.
Once you have selected your charge weight, do a seating depth test in .003" increments between something like .006" and .024" off the lands. I typically start charge weight testing at .015" off the lands, which is exactly in the center of this seating depth range. That way, the bullet won't be moved more than half the total seating depth range in either direction from the initial charge weight test seating depth. I have not found that seating depth changes that small affect velocity, pressure, or relationship to OBT nodes to any significant degree. The seating depth test is where you want to tighten up your groupings. Look for two to three .003" increments in a row that give the smallest groupings, then choose the longest CBTO of these. This will give you the most headroom for future land erosion.
Bottom line is that there are many different ways to end up at the same place in load development, and everyone has their own personal favorite method. However, this is ballistic science and I treat it as such. Testing changing ONE variable at a time in SMALL increments is how you do that. Then you know with certainty whether some specific parameter you changed had an effect.
Prior to using Quickload, I had developed many optimized loads through more "standard" approaches such as OCW and/or ladder tests that I later found ended very close to an OBT node (just a slight tick slower as I mentioned above). Now all I do is shoot 5-10 rounds for a new load/rifle at a charge weight predicted by QL to be slightly below where I think the OBT node will be for a given powder/bullet and measure velocity. I plug the rifle/load specific variables into QL and adjust the preset powder burn rate until predicted velocity exactly matches actual velocity, in effect "calibrating" QL for that specific setup. Then I set up a charge weight test in 0.1 gr (fine) increments with the predicted OBT node in the middle as I suggested for you to do above. I have yet to do one of these charge weight tests where low ES/SD corresponding to an OBT node didn't fall within in the 0.1 gr increment charge weight test window that was predicted using "calibrated" QL. In my hands, these loads typically also produce the smallest groups. Although the key is to find the optimal charge weight window that produces low ES/SD (as opposed to the smallest groups), it's comforting to usually see the groups shrink in that region as well. I load to the middle of the window, optimize seating depth (all at 100 yd), then validate the load under simulated match conditions (typically at 100 yd also, but sometimes at 300 or 600 yd). Using QL can save a significant amount of time and reloading components, and is a very useful investment, IMO.
In any event, I'm guessing based on the predictions that somewhere between 24.5 and 24.9 gr will be your optimal charge weight, likely 24.6-24.7 gr. Let us know what you find. Good luck.
