50-45, 50-45, 47-48
The odd decision of Felder/dos Anjos surfaced some questions on how to handle split decisions where there is one super obvious outlier. Do you respect that all judges have an equal say? Do you take the one in the middle (median)? Something else?
Let’s take a look at the effects of Felder’s rating in the various systems — dos Anjos switched divisions, while Felder remained at 155, which makes him a better subject for this within the standard system.
| System | Win Weights | Pre-Fight | Post-Fight | Diff |
| Fight Matrix “Standard” | 72 (90,97) | 427 | 427 | = |
| ELO “K170” | 100 (100,100) | 1909 | 1828 | -81 |
| ELO “Modified” | 67 (100,100) | 1929 | 1894 | -35 |
| Glicko-1 | 55 (91,91) | 1985 | 1967 | -18 |
| WHR | 69 (88,99) | 1839 | 1819 | -20 |
The win weights list three weights (the “power” of a win out of 100) for:
- The actual fight.
- The fight if it were scored 49-46 x3 UD (average).
- The fight if it were scored 50-45 x3 UD (majority & median).
In all systems but the standard, Felder lost points. OK, if we are to get technical, Felder lost 0.0051 points in the standard system, but that had nothing to do with the fight itself. The standard and the WHR systems were the least affected by this egregious error.
