Firts, I arranged a white panel and a black panel so that they split the frame exactly in two. These images are unedited, directly out of my D3, low ambient conditions, f/5.6, 1/200th, ISO 400, regular TTL.
Image 1) In this image, the focus point was moved to the far left in the center of the white panel (the panel appears gray, because the flash system has averaged the center weighted frame and chosen a power that balances the entire image to 18% gray.
Image 2) In this image, I moved the focus spot all the way to the right side in the middle of the black panel. Notice that this image is exactly the same exposure as Image 1) showing that the position of the focus point is not used for the regular TTL metering system.
In the next few images, I investigate the effects of the focus system on TTL compared to TTL-BL.
Image 3) In this image the flash was in TTL and the subject was placed in the left part of the frame and the leftmost focus point selected. You can see that the statue is slightly overexposed from being off-center, but that is not what I am studying this time.
Image 4) In the next image, (still in TTL mode) I moved the focus point to the far right on the rope extending below the hats, and allowed the camera to focus there. You can see that moving the focus point did not affect the flash metering. The subject (although now out of focus) is exactly the same brightness as the previous image, as is the background.
Image 5) In the next image I switched to TTL-BL mode and placed the focus point back on the statue. You can see that the statue is no longer overexposed, as the balancing equations try to balance it with the darker background.
Image 6) In the next image I remained in TTL-BL mode and moved the focus point to the far right so the rope below the hats would be in focus. Immediately you see a change in the flash power. The statue is now overexposed and the background is much brighter. This is what has been causing even experienced photographers to think that the focus selection has an effect on flash power. However, the change is not being caused by moving the focus points. It is being caused by the change in focus distance. In this image the hats are about 15 feet behind the statue, and in TTL-BL mode, the distance is the prime factor that is used to set flash power. So, the farther distance caused the power to increase.
Image 7) In the next image I left the flash in TTL-BL mode, and switched off the auto focus and moved the focus point back to the statue. the important thing to notice is that the exposure of this image is identical to the previous image. This proves that moving the focus point does not affect flash power.
Image 8) In this image I moved the focus point back to the hats. Again the exposure did not change.
Image 9) In this image, I moved the statue into the center of the frame and selected regular TTL. Now the exposure of the statue is correct (the color is now accurate as well). This shows that even with all of the advancements of how TTL-BL deals with a dark background, regular TTL is still better as long as the subject is centered properly in the frame and occupies about 20% of the frame so the monitor preflashes can meter it accurately.
These images prove that moving the focus point does not affect flash metering in either TTL or TTL-BL flash modes. The change in exposure that has been observed in TTL-BL mode when moving focus points is due entirely to the change in focus distance as reported by the D lens.