My guess would be that images with equal samples should have equal influence (transparency)

E.g. for five images: (click for full size)

Each image is rendered with 10 samples and is combined with an alpha of 20%.

First image:

Combined result:

##Combining with an image with more samples Since the output of each Alpha Over operation results in a image that is completely opaque, we can forget about the previous operations when combining the result with more images and treat the combined image as a render with ~images*samples per image samples. Basically all we need to do is add another alpha over node and set the Factor based on the number of samples of each image.

To combine the above stacked image with an image rendered with 50 samples we should use a factor of .5 because the total combined samples of the stacked images is approximately equal to 50 as well:

Straight 50 sample render:

5 combined 10 sample renders combined with the straight 50 sample render:

Short answer:

My guess is that it should be 50%. The combined samples of the images above it is ~750, which is the same as the number of samples in the rendered image. Since the rendered image and the combined images are more or less equivalent, They should be combined with a 50/50 mix.

Long answer:

Each image's opacity is based on how many samples it has compared to the image above it.

E.g. if you render 5 images with 10 samples each, the second image should be mixed with the first with a factor of .5 (or an alpha of 50%). This will effectively result in a 20 sample image (as you pointed out, it's not quite the same as 20 rendered samples, but it's reasonably close), so the second image should be mixed with a factor of .33 (alpha of 33%) and so on.

Since the resulting combined image is basically a 750 sample render, it's the same number of samples as your 750 sample render. You can just tack it on top with a transparency of 50%.

My guess would be that images with equal samples should have equal influence (transparency)

E.g. for five images: (click for full size)

Each image is rendered with 10 samples and is combined with an alpha of 20%.

First image:

Combined result:

##Combining with an image with more samples Since the output of each Alpha Over operation results in a image that is completely opaque, we can forget about the previous operations when combining the result with more images. Basically all we need to do is add another alpha over node and set the Factor based on the number samples of each image.

To combine the above stacked image with an image rendered with 50 samples we should use a factor of .5 because the total combined samples of the stacked images is approximately equal to 50 as well:

Straight 50 sample render:

5 combined 10 sample renders combined with the straight 50 sample render:

My guess would be that images with equal samples should have equal influence (transparency)

E.g. for five images: (click for full size)

Each image is rendered with 10 samples and is combined with an alpha of 20%.

First image:

Combined result:

##Combining with an image with more samples Since the output of each Alpha Over operation results in a image that is completely opaque, we can forget about the previous operations when combining the result with more images and treat the combined image as a render with ~images*samples per image samples. Basically all we need to do is add another alpha over node and set the Factor based on the number of samples of each image.

To combine the above stacked image with an image rendered with 50 samples we should use a factor of .5 because the total combined samples of the stacked images is approximately equal to 50 as well:

Straight 50 sample render:

5 combined 10 sample renders combined with the straight 50 sample render: