Player Usage Error?

The new Player Usage announcement mentions "How often a player was used" but the directory states that is is a rating of performance. Anyone know which one is correct?

Within the Stats History tab of the Player Set Browser appears to be a new usage rating, where the cards that have been used the most in 80 million auto leagues get a rating of 10 and it scales downward from there.

The new usage rating is definitely not based on performance, as there are several players I've checked whose .50 card is a 9 or 10 and their more expensive card is considerably lower.

It certainly appears to favor cards that have been around longer. Charles Gipson gets a 10, for example, though he hasn't been used nearly as much since his price went from .50 to .77. And Luis Arroyo's more expensive card is a 10, but he isn't used anywhere near as much in ATG 9, after the bullpen rule change.

It's certainly very interesting to look at. I have a feeling the weights are heavy, so if a single manager uses a card once or twice it has a surprisingly large impact on the rating.

Not sure about that. I've used Austin Meadows card a bunch at 80m and he is only a 3.

Not sure about that. I've used Austin Meadows card a bunch at 80m and he is only a 3.

That is surprising but good to know. Thanks.

It's a fantastic and long-overdue addition. It shouldn't be tracking usage across price changes, it claims to only track usage within a player set, and the numbers indeed change if you select an earlier ATG. Or even ATG9 1901+.

It's a fantastic and long-overdue addition. It shouldn't be tracking usage across price changes, it claims to only track usage within a player set, and the numbers indeed change if you select an earlier ATG. Or even ATG9 1901+.

Now that I've had more time to check it out, that does appear to be the case. Looking back at ATG8, the old favorites are all 9 or 10, with some significant drops in ATG9. I guess it's just me who isn't using Gipson or Arroyo much anymore!

Then again, Bruce Sutter only dropping from a 10 in ATG8 to an 8 in ATG9 is hard to believe.

It's a fantastic and long-overdue addition. It shouldn't be tracking usage across price changes, it claims to only track usage within a player set, and the numbers indeed change if you select an earlier ATG. Or even ATG9 1901+.

Now that I've had more time to check it out, that does appear to be the case. Looking back at ATG8, the old favorites are all 9 or 10, with some significant drops in ATG9. I guess it's just me who isn't using Gipson or Arroyo much anymore!

Then again, Bruce Sutter only dropping from a 10 in ATG8 to an 8 in ATG9 is hard to believe.

I found Sutter's rating to be high too. I have only seen him once since I came back to the game iirc. Perhaps I'm forgetting some but an 8 sounds high. However, I don't play auto leagues, I just play DH leagues. So with more money to spend on pitching I could see it. Speaking of that, it would be cool if this feature existed for DH leagues too.

One thing that could really put this feature over the top is if it's smart enough to account for new cards. I.e., the usage rate for each card should be based on the number of leagues created since the card was added. Otherwise new cards will appear less popular than they are.

I plugged in some popular new cards and several of them are listed as 10s. But the last addition was so long ago, they may have just had enough time to enter the top 10% despite being new.

Upon further review, the thing that jumps out at me is 9s and 8s on guys who I never have any problem acquiring. I think what's going on is that the vast majority of cards cluster around 0% usage across leagues. So when you display their usage rate by percentile, 1s are used in 0.0% of leagues, 2s are used in 0.1% of leagues, etc. Meanwhile 10s are everything above 30% or so. 9s are about 15-30%, 8s are about 7-15%, 7s are about 5-7%. Sooo, although this is better than the status quo ante of no information about card popularity whatsoever, presenting the info in terms of percentile really seems to limit the usefulness.