When you don't understand something, you seem to just conveniently ignore it. Why don't you go back and field the last question put to you in #249 regarding your understanding of a figure-of-merit? I put in the effort to explain what you failed to understand then, but here I am still faced with criticisms emanating from the same ignorance.
I
did respond, by refusing to comply with your anally retentive standard, and explaining why there's little use in performing the meticulous work necessary to prove something like this numerically. Just because I didn't give you the kind of answer you wanted doesn't mean I'm "ignoring" you. Remember this:
Well, dude, it's a complicated game. You're just not going to get anything much more specific than that. If that's what it takes to prove to you that a card is "good," then pretty much all cards are going to be mysterious in their utility to you.
Now, granted, I am in accord with the methodology that's been presented since. However, I feel that it would require a very large sample of matches to be of any value because the case in which the card makes the difference between a win and a loss is certainly rarer than the much more difficult to measure case where it simply helps the winning effort. I voiced these concerns here:
This whole concept seems too limited, only taking into accounts the games where it makes the difference between winning and losing. It can still make your win easier, faster, or less risky, even if you were going to win anyway, and that would still be "helpful" and should still probably count towards its favor. Not that that's objectively observable in any remotely simple way.
...to which you responded with something totally irrelevant, behaving as a jerk and hypocrite at the same time. You most certainly did not "address my point," no more than I did yours when you decided to criticize me for being sarcastic and unprepared.
Now to actually return to the topic at hand.
Like I've pointed out, I feel that many, many matches have to be recorded for the resulting figures to have any meaning. All of this just to satisfy one person's demand for a numerical approach? I'm surprised at the diligence of those that, in part, agree with me here. What I'm saying is that all of this is really not worth it to prove the utility of a single card, which is why it's rarely done (not never, though). Theoretical arguments, of which the rest of the thread has consisted, are not completely worthless, and you shouldn't ignore them as a basis for a position. Of course, they're not as concrete as a well-maintained practical study like what has currently begun, but they can still be convincing and can still make logical sense. Your refusal to accept anything less shows either closedmindedness or major **** retention, and your willingness to completely overlook such arguments suggests it is the former.
I don't even want an adequate response from you to the post that really started this. In retrospect, it was not as thought-out as it should have been. The limitation that concerned me can be overcome by a large enough sample, but that entails its own challenges. If you want to respond to
that, be my guest. I can't imagine a negative response to it other than the assertion that a large sample is
not needed (with reasoning, I would hope), and such an argument is something I would be glad to hear.