6.6.1 Occasional failure

It suffices for the disclosure of an invention that the means intended to carry out the invention are clearly disclosed in technical terms which render them implementable and that the intended result is achieved at least in some, equally realistic, cases (T 487/91). The occasional failure of a process as claimed does not impair its reproducibility if only a few attempts are required to transform failure into success, provided that these attempts are kept within reasonable bounds and do not require an inventive step (T 931/91). The skilled person is used to occasional failures when testing a technical teaching (T 14/83, cited in T 1133/08).

In ex parte case T 1111/14 (human hepatocytes), the board, disagreeing with the department of first instance's finding on Art. 83 EPC (infringed), cited the above finding from T 931/91 and added that, as a matter of fact, occasional failure was the rule rather than the exception in the technical field at issue.

A claim is an attempt to define a device in terms of ideal conditions, i.e. those required for its theoretically optimal or nominal operation. However, when considering a claim, the skilled person will readily understand that the conditions of actual operation will not be the ideal ones defined there. In T 383/14 (sorting table for grape harvest), the board found that, on reading the claim at issue, the skilled person would immediately grasp how the table would operate in practice after a harvest and so understand its terms in a sense compatible with the actual operation of all mechanical devices, whose reliability or success rate was always less than 100% and even lower in the specific case of sorting or grading.

The board in T 38/11 summarised the case law for making a case of insufficiency of disclosure (identifying gaps in information), and in the case at issue stated that the appellant (patentee) itself argued that a synergistic effect of a composition depended on a range of parameters and was rather an exceptional situation. As such parameters were not disclosed, it followed that the patent did not suffer from an occasional failure, but from a lack of a concept fit for generalisation. The situation may be aptly denoted as an invitation to carry out a research programme, based on trial and error, with limited chances of success (see T 435/91 (OJ 1995, 188) and T 809/07). In accordance sufficiency of disclosure could not be acknowledged.

In T 2728/16 the board concluded that the patent did not disclose a technical concept fit for generalisation, thus the majority of the systems encompassed by claim 1 were not available to the skilled person. A teaching of how the non-functional embodiments could be made functional was not at the disposal of the person skilled in the art even after reading the patent. Thus, the non-functional systems of claim 1 were not the result of an occasional failure that could be tolerated and turned into a success without undue burden, but of a systematic occurrence. The decisions T 487/91, T 931/91 and T 107/91, cited by the respondent (patent proprietor), were thus not applicable. The board did not agree with the respondent that the reported failure rate of 2% of the experimental runs meant that 98% of the systems covered by claim 1 provided meaningful results. The experimental runs described in the patent did not encompass substantially all embodiments falling under claim 1.

In T 603/22 (method for milling PAEK scales to obtain a powder having an average diameter of less than 100 μm), the board held that the failures in this case were not "occasional" since the patent did not show that the claimed results were achieved in at least some cases (see also T 487/91). This was in fact a case of systematic failure. The appellant (patent proprietor) also relied on the case law to maintain that the necessary trial and error was acceptable. The board held that the case law stipulated limits on trial and error (see the section below in this chapter II.C.6.7.).