The "Common successor"-heuristic ist described here: Motamedi et al.: A Survey of Techniques for Internet Topology Discovery, p. 11, 2015 http://ieeexplore.ieee.org/abstract/document/6970764/
From the above named source:
The “common successor” heuristic suggests which two IP addresses may be aliases. This heuristic relies on the prevalence of routers that respond to traceroute probes with the incoming interface. When two traceroute paths merge, the common IP belongs to the second router on the shared path. IP addresses prior to the common IP should belong to different interfaces of a single router and hence would be aliases. Figure 7 shows a partial view of a traceroute measurement from Host1 and Host2 toward Host3 in our toy example. In this example, the black interface succeeds the red interface in one traceroute and succeeds the blue interface in another traceroute. The heuristic suggests that the blue and the red interfaces are aliases. This heuristic falsely infers aliases in the presence of layer 2 switches or multiple-access clouds. Figure 8 depicts an alternate topology to Figure 7. The traceroute view in both figures are similar, hence the heuristic infers R26’s red interface and R23’s blue interface are aliases.
What I do not understand: Even if there was no layer 2 switch and R23 and R26 would just route to R21, wouldn't they be regocnized as aliases by this heuristic. I just do not get how you can say two routers must be the same, because the have a common successor. Maybe the paths just start to merge in the "successor".