Not sure if there was supposed to be a link or quote explaining the 1/400 rule of thumb. It is found in some infrared photography books and suggests increasing the extension by 1/400 of the focal length (effectively that the focal length is longer by 1/400, or 0.25%). This was intended to apply for Kodak HIE type IR film and the correction for current IR films less extended into the IR is probably less.
I once took several lenses with a manufacturer's IR focus mark - typical 1980s 35mm SLR lenses - and compared the amount of extension (calculated from the focus distance) to the 1/400 rule of thumb. It's in the right ballpark, perhaps a little more than the manufacturer's estimate on average.
For a front element focusing lens, the amount of physical extension of the front element is not exactly equal to the change in focal length. It may be better to compute the focus scale mark corresponding to the "IR infinity." For example, if we think the IR focal length should be 110.275mm, that corresponds to an optical focused distance of:
1/(object distance) + 1/(effective lens distance) = 1/(focal length)
1/(object distance) + 1/110.275mm = 1/110mm
so distance = 44 meters. So you would put the IR focus mark opposite 44 meters on the distance scale.
This isn't much of a change, from the infinity mark to 44 meters. For ex, with a circle of confusion of 0.06mm, which is about right for a 6x9 folder, a focus shift of 0.275mm is equal to the depth of focus at f/number = 0.275mm/0.06mm = f/4.6. Thus, even with the lens wide open at f/4.5, the one-sided depth of focus is just covering the predicted focus shift for HIE wavelengths. For a film like Rollei IR, the focus shift is likely less, which is probably why some people report ok results without shifting the focus at all. Again, these are all rules of thumb, and there might be some lens design out there for which the focus shift is larger depending on the specific glasses used. But a triplet is fairly common, non-exotic.