(A) Example simulation for the simple scenario with informative test choice (SIM-1). The correct sequence is ABC. In the first round, all hypotheses have the same prior probability of 1/6, and all tests have the same informativity. One test, BA, is chosen randomly and yields a positive. Since BA is not part of sequence ABC, this is a false positive. The probabilities for ABC and three other sequences decline, while the probabilities for the two sequences that contain BA (BAC and CBA) increase. In the next round, the tests AC and CB are the most informative ones. They distinguish between the two most likely hypotheses, BAC and CBA. AC is chosen and yields a negative result (true negative). This weakens hypothesis BAC and supports CBA. In the third round, CB is the most informative test. A negative result is obtained, and CBA and BAC are on par again. Further tests are performed, and yield correct answers, which establishes the correct sequence ABC as the most likely one. However, in the last round AB is tested and yields a false negative. The probability for ABC declines and finishes on par with BCA. (B) Example simulation for the complex scenario where in each round, two tests are performed but only one test results can be published. Again, ABC is the correct sequence. In first round, where all tests have the same informativity, two tests are chosen randomly. Both tests BA and AC yield a negative result and turn out to be equally valuable. The negative result on AC is randomly chosen to be published. This decreases the probabilities for ACB and BAC, and increases the probability for the four other sequences. In the second round, AB and BC are tested. Both tests yield false negatives, one of which (BC) is published. This leads to a decline for the probabilities of ABC and BCA. After a few rounds of testing, CAB is leading while the correct hypothesis ABC is second best. However, a true negative on CA bring CAB and ABC on par, and further tests establish ABC as the most likely sequence. In both panels, italic type codes for false positives and negatives.

## PubMed Commons