Posted by grappler0000
Posted by lohmann
There's a lot of problems with FightMetric...
Could you elaborate on the problems?
FightMetric uses an algorithm to determine their own effectiveness score, sometimes at the expense of what the criteria for judging is.
For grappling:Grappling actions have equal opposite reactions - The same number of points granted to one fighter for a position change (e.g., gaining mount) must be awarded to the other fighter should he work his way out of it.
This particular rule ignores the way judging is handled and instead inserts something totally different. Offense that leads people to move into more dominant positions carries more weight than any defensive maneuver used to counter.It's what happens after the takedown that matters - With the exception of slams, takedowns are only valuable insofar as they set up more valuable opportunities on the ground. Taking someone down into guard confers only a modest advantage to the fighter in top position.
If two fighters are on the ground for a duration of a round and neither holds a significant edge in the amount of damage given, the fighter with a dominant position is awarded the round according to judging criteria because strikes thrown from the top are weighed more than those from the bottom. That's more than just a "modest advantage." Dominant positions matter much, much more, even if I would prefer the way FightMetric sees it to be the way judges do. Henderson/Cerrone and Lawal/Mousasi are two recent examples where FightMetric awarded rounds to fighters that may not have deserved a round merely because they place an emphasis on damage instead of an emphasis on control.
FightMetric inputs data in raw numbers, which can oftentimes fail to tell the reality of a fight considering the infinite complexity of a single strike to be accounted for properly via any algorithm. FightMetric scores every single strike thrown by a fighter in some way, including soft blows in a very close guard, when judges should and mostly do find those sort of strikes nearly negligible.
If you have not seen the Compustrike numbers
for Jung vs. Garcia, you see raw numbers collected in a different and inferior way that tells a different type of story about a fight that should, in my opinion, not be as close as the fight was seen.
In their own FAQ: The problem with current judging systems is that no one knows what techniques to value more than others. Perhaps the FightMetric system and its underlying data can inform that discussion.
Judging is subjective and often disappointing. FightMetric is the one website I trust more than others, but they are not free from error either.
The one time I disagreed with FightMetric that I can ever recall was Evans vs. Silva
, where Evans demonstrated an amount of control that is difficult to process as information and should have, doubtlessly, been awarded the first and second rounds.