For the singles short program jump elements there is a proposed maximum possible credit for jump elements of 3.0 for the three jump elements (Axel jump, solo jump out of footwork, jump combination). If we assume each jump element is of equal importance to the requirements of the short program, than each element has a maximum potential credit of 1.0.
If all three elements are completed, then the sub-mark for jumps would be based on the difficulty of each jump element and the quality with which it was executed. To receive the maximum mark of 3.0 the skater would have to execute the most difficult conceivable example of each required jump element perfectly. Jumps of lesser difficulty or lesser quality would receive less than a 3.0 sub-mark.
If one jump element is omitted or complete failed, then the maximum possible score for the sub-mark for jumps would be 2.0. To receive this maximum mark the skater would have to perfectly execute the most difficult conceivable examples of the two elements completed. Jumps of lesser difficulty or lesser quality would receive less than a 2.0 sub-mark.
The omission or complete failure of two jump elements would be similarly handled, with a maximum sub-mark for jumps of 1.0. The sub-mark for jumps would be 0.0 if all three jump elements were omitted or completely failed.
Consider now the two top men attempting short programs with 4T3T, 3A and 3L. One skater completes all three elements, while the other falls on the 4T and his combination is marked a complete failure. Under the current 6.0 system, both skaters would receive similar base marks in the high 5's (based on reputation) and the skater who fell would receive a deduction of 0.4. At Worlds, the skater who fell would typically end up dropping four places compared to the leader due to this error and would "need help" to win overall.
Under this proposal, the skater who fell could only receive a maximum possible Technical Merit mark of 5.0, which is a fairly low score under the current 6.0 system. However, under the proposed revision to the meaning of a 6.0, the skater who landed all three elements would only have received a Technical Merit mark of about 5.2-5.4. The skater who fell would receive a Technical Merit mark of 4.8-5.0. At worst, the skater who fell would trail the leader by 0.6 in the Short Program, but after taking into account the 50% weight of the short program, would only have a deficit of 0.3 to overcome in the long program. Something that would be reasonably achievable, with no help required.
In this example, the skater who fell, though he only received a Technical Merit mark of 4.8, would still end up near the top in the Short Program, and still would have a chance to win the event in the long. The skater would also not be bypassed by lesser skaters who attempt elements of significantly lesser difficulty, because under the revised meaning of a 6.0, those skaters in the middle of the pack would now receive marks in the low 4's or less, while poor skaters would receive marks in the 3's or less.
Although the Technical Merit scores would change somewhat in numeric value compared to the current 6.0 system, in this approach the relative penalties and placements would remain similar to the past, with the important exception that a top skater who fell on one element in the short program still has a chance of winning the event with a strong skate in the long program, with no help required from other competitors.
In addition to eliminating deductions for omissions and complete failures, which would now be handled in the base mark, all other deductions that could be handled directly in the base mark would be greatly reduced or eliminated. For example, if a triple jump is required in the solo jump and a skater executes a double, a substantial deduction is not required since the reduced value of a double compared to a triple can be directly accounted for in the base mark. The only reason to have the deduction for insufficient rotations would be to "motivate" the skater to attempt the required triple instead of taking the easy way out with the double. This could be handled with the 0.2 deduction for an element being "not according to requirements."
For errors that are true defects in the execution of an element, the deductions would be retained and serve the purpose of specifying how much credit an element should lose due to the defect. Errors such as slightly cheating a jump, a toe touch, step out, hand down, etc. are examples of true defects in an element -- as compared to doing a double instead of a triple which is an example of executing a less difficult element in place of a required more difficult element.
In summary, "errors" that consist of doing an element of lesser difficulty than required would be handled by a reduced base mark and a Not According to Requirements deduction, while errors that are true defects in execution would be handled in the base value for the element actually attempted and a deduction for the specific defect in the attempt of that element.