shifting at 7k means you drop to around 3800 rpms on the 1-2 shift and 4500 rpms on the 2-3 shift. 2nd gear gives you more torque multiplication than 3rd gear and makes up for the lower rpms after the shift much quicker than 3rd. With the exception of the initial hit, there is almost zero torque multiplication from 3rd. With a delayed lockup in third, the converter is slipping (a lot) and that means the engine will accelerate faster but that does not necessarily mean the car is accelerating faster.
The best shift point is defined in drag racing as the least loss of acceleration after the shift. Since acceleration is a product of torque and rpm, you can calculate the acceleration from the dyno plot in rough intervals to arrive upon the optimum shift point - realizing these values are with the converter locked.
The wild card is the torque converter lockup delay. With the torque output assumed flat, the best acceleration is obviously with higher rpms. The trade off to get higher rpms is the delay in lockup. If you delay it too much however, the overall acceleration will be lower since the total acceleration is the sum of the unlocked and locked acceleration (remember the tortoise and the hare story?).
Furthermore, each converter will have a different output unlocked than locked and it will vary with rpms. To 'copycat' one setup to another is not going to guarantee the same results.
Long story short... Only by comparing ets at the track with different set points can you optimize the settings. You may
think the car is faster with a higher shift point and a more delayed lockup but your et may actually be slower. Likewise, you may feel the car is 'hitting the wall' with an earlier lockup but your et may actually be quicker.