The weighing of evidence involves the connecting of dots, which involves drawing inferences. However, just as there can be false positives, there can be false inferences. The tendency may be to think of inferences as coming in one size, but not all inferences are logically correct. Some are weak and others strong. The reason for talking about both kinds is that so much depends on the quality of their making. One party presents facts supported by circumstantial evidence that the opposing party rebuts with other facts; or not, if respondent defaults. From that brew there is no avoiding of inferences. It is a normal cognitive process. It would be difficult without them to form opinions in the absence of any probative evidence, which does not mean that opinions are not formed from weak inferences. That is why in a well-reasoned legal decision courts sort out the facts and consider the right inferences before they announce their decisions.

In the best reasoned decisions under the Uniform Domain Name Dispute Resolution Policy (UDRP) Panels are equally concerned with the path taken to conclusion; Examiners under the Uniform Rapid Suspension System (URS) do less reasoning (and, in many cases, none) because of the nature of the proceeding, which is only applicable to “the most clear-cut cases of infringement.” Even if the word “inference” was not explicitly authorized under the UDRP at least for default, Rule 14(b) (Default)), it would still be used by decision-makers. Rule 12(f) of the URS expands the drawing of inferences from default (Rule 5(i)) to noncompliance “with any provision of or requirement under, these Rules, the URS Procedure or the Provider’s Supplemental Rules.” This general authority beyond default is not written into the UDRP but supported by twenty years of Panel construction.

False inferences happen when Panels assume too much from or put too much weight on any one piece of supposed evidence: the decision in Autobuses de Oriente ADO, S.A. de C.V. v. Private Registration / Francois Carrillo, D2017-1661 (WIPO February 1, 2018) (<ado.com>) decision (discussed in an earlier essay) falls into this category. Unless strings of letters, dictionary words alone or combined, generic phrases, or common expressions have such high recognition in the marketplace that denying targeting would amount to willful blindness, liability should not rest on superficialities such as “Respondent’s position as a professional domainer,” particularly if the letters in issue are used as commercial marks by many other businesses as (incidentally) was the case with “a,” “d,” and “o.” A similarly false inference was made in Irving Materials, Inc. v. Black, Jeff / PartnerVision Ventures, FA1710001753342 (Forum November 7, 2017) involving a three-letter string, <imi.com>, acquired over twenty years prior to the complaint. Both Respondents are presently challenging the awards in federal court under the Anticybersquatting Consumer Protection Act (ACPA). Respondent in <ado.com> was represented and put in its evidence, but Respondent in <imi.com> did not appear and the Panel interpreted the evidence (such as it was) against it.

In calling attention to false inferences, I am not suggesting they are endemic; far from it. But parties should be careful what goes into the record and Panels (UDRP) and Examiners (URS) careful what they read into allegations and thin proof. (Admittedly, they are working under extreme time pressure (14 and 3 days, respectively). Nevertheless, they should avoid dubious application of willful blindness, allegations and assumptions, and unexplained circumstances as substitutes for evidence.

Three recent examples from the URS docket are useful as introduction. I will then review a couple of recent UDRP decisions in which Panels explain what it means to draw inferences. Parties and their representatives should pay attention to the reasoning process for pitfalls in their submissions.

In The Boston Consulting Group, Inc. v. yanmingcui, FA181100 1815095 (Forum November 26, 2018), the Examiner held that “[g]iven the longstanding worldwide use by Complainant of the BCG mark, it may also be assumed that the domain name was registered in bad faith” (Emphasis added). The rule is, “nothing can be assumed.” Assumptions are not probative. While it may be true that a complainant may have many trademarks globally, in the case of “bcg” Complainant is only one of many other companies using the same three letters. In The Bureau of National Affairs, Inc. v. FA1811001815436 (Forum November 27, 2018) (<blaw.space>) the Examiner does not even explain how the four character string could be identical or confusingly similar to the registered acronym BNA (No mention is made about the BLOOMBERG LAW mark, although that is what the Examiner may have had in mind).

Three days after <blaw.space>, the Examiner in Bureau of National Affairs, Inc. v. Claim Number: FA1811001815433 (Forum November 30, 2018) (<bna.ooo) illustrates how making inferences should properly be done:

Although this Panel recognizes that Complaint’s mark has been extensively used over decades and enjoys considerable reputation and fame among the relevant public, there is no evidence on the record clearly showing that Respondent registered its domain aiming to profit from its reputation and goodwill, particularly in view of the following. BNA is an abbreviation standing for a whole number of shortened words or phrases. Also, Complainant is active in a specific niche sector of news services focusing on law, tax and environment. Given this, proving bad faith under the circumstances of this case puts a higher burden of proof on Complainant, hence making a decision in this case within the URS procedure may be not possible. Therefore, it is found that Complainant has not proved that Respondent acted in bad faith when registering and using the domain name by clear and convincing evidence.

At the opening of any litigable/arbitral dispute, there is a mixed brew of alleged facts, some supported others not, and where supported parties may offer either direct or circumstantial evidence tending to prove the truth of the facts they rely on. When the facts come from one party only, say complainant because respondent has defaulted, it is more likely than not complainant’s facts will control the decision, but the more attenuated or debatable the evidence, the more the alleged facts have to be weighed carefully in an unbiased manner.

Where allegations and inferences proposed or drawn are supported by unrebutted documentary evidence they can be accepted as resting on a logical base. To take an example: in Dominique Tillen v. Administrator Administrator, D2018-2181 (WIPO November 12, 2018) (<brush-baby.com>, not exactly a common phrase, thus an issue the Respondent must address) the parties had been negotiating a distribution agreement when Respondent registered the domain name but the negotiations did not proceed further than a draft memorandum of understanding:

The Complainant has provided an MOU regarding this distribution arrangement, which refers to a full distribution agreement which was yet to be drafted. The MOU provided in the Complaint is not signed by either party. The Complainant states that the distribution arrangement did not proceed. As such, the Panel assumes [that is, infers] that the full distribution agreement was never prepared.

This naturally leads to the conclusion (not based on supposition but on direct evidence of a fact) that “the preparation of an MOU did not give the Respondent the right to register the Disputed Domain Name.” Further, “[e]ven if the MOU had been signed, it would be open for the Panel to consider that the Respondent only had permission to use the Trade Mark (including in the Disputed Domain Name) during the term of the distribution arrangement.”

Where there are allegations of fact but no follow through with evidence, no inference can or should be drawn. Complainant in Zoyo Capital Limited v. A. Zoyo, D2018-2234 (WIPO November 13, 2018) () alleged bad faith on the grounds that “[o]ffering of the disputed domain name at an inflated price suggests the Respondent’s knowledge of the business value of the disputed domain name and that the Respondent intended to make commercial gain from it,” which is not the law. The Panel concluded that

On the face of the record the Respondent’s family name is Zoyo. If this is correct, then prima facie, the Respondent may have a legitimate interest in the disputed domain name. The Complainant has provided no evidence suggesting that the Respondent’s name is fictitious. In these circumstances whereby the Complainant is silent as to this issue and absent any indication in the record that the name is false, notwithstanding that the Respondent has failed to rebut the Complainant’s contention, the Panel has difficulty in concluding that the Complainant has made out a prima facie case that the Respondent has no rights or legitimate interests in the disputed domain name.

(Note: Although complainants and their representatives should not count on it, Zoyo is one of several cases that have allowed deficiencies of proof to be corrected by authorizing complainants to refile complaints, https://wp.me/p5aTJ6-1kn).

Success comes through preparation and anticipation of the impact of the submission on the decision maker. What goes into the pleadings and proof, after all, determines what conclusions decision makers draw from them. I am suggesting that parties should also pay attention to the grades of inferences that can be drawn from their submissions and omissions and to learn from decisions what more (or less) should be done to persuade decision-makers in their favor. Complainants prevail by building structures of evidence that logically progress to the conclusions they want decision makers to reach (Dominique Tillen understood and Toyo Capital Limited did not). Respondents prevail either by offering affirmative evidence of their rights or legitimate interests or by poking holes in the evidence of bad faith. I have also suggested that decision-makers can get it wrong by assuming unlawful registrations of domain names where the record supports no such intention.