If you use the multi-bitrate encoding in the MS Expression Encoder and encode the files in the same session from the same source then yes they should be chunked the same. If the files are from different sessions or use a different source then Expression encoder might be using screen frame detection to insert key frames at scene changes and this might cause a variation between files.
I am not sure what you think is wrong. If I open all four renditions they are chunked exactly the same. All four renditions are broken up into chunks exactly the same way. If you are expecting each rendition's chunk size to be exactly 10 seconds this is not going to happen unless the key frame interval is perfectly broken on values that match the chunk duration. If you make the key frame interval a factor of the chunk size such as 2, 5 or 10 then this might work. Then again it might not lead to prefect chunk sizes because the encoder might inject key frames off the key frame interval due to scene change detection or the key frame interval might be off by one frame due to how it interprets the interval. Either way it doesn't matter. What is important is that the four files are broken up the same way.