Metengr is correct that there are really too many variables to come up with any good correlation. However, with high strength Q&T steels, I've found the YS/UTS ratio to correlate well with the amount of martensite formed on quenching; the higher the ratio, the higher the amount of martensite. Since, in general, the more martensite formed on quenching the higher the impact strength, there is a correlation. You do have to limit the varriables, particularly chemistry. But, for two Q&T samples of the same steel grade at the same tensile strength levels, the one with the higher YS/UTS ratio will generally have higher impact strengths and lower transition temperatures. Of course, there are other factors to consider, but if you control these other factors, you can find a decent correlation. For example, with 4140 Q&T to 30-35 HRC, if you have a YS/UTS ratio above 0.85 or so, I'd expect impact strengths around 50-80 Joules (40-60 Ft-Lbs). At the same hardenss, but YS/UTS ratios below 0.80, impact values will be more in the 20-40 Joule (15-30 Ft-Lb) range. Like metengr, this observation is only based on my expereince and I really don't have any specific explanaiton for this behavior.
I find it very interesting that this is in direct oppositon to metengr's observations. Although my experience with mechanical properties of normalized steel is low, I understand and agree with his comments and explanations. I think the differences result from dealing with a ferritic/pearlitic microstructure as opposed to a martensitic/bainitic microstructure.
rp