| |
BGonline.org Forums
Rollout eXtension
Posted By: Nack Ballard In Response To: Rollout eXtension (Stick)
Date: Sunday, 20 June 2010, at 11:09 p.m.
Like I said, you really have to devalue the rollout when done by one CB.
54S-21H-31
Perhaps Chuck was just seeing if he could goad you into extending. :)
So, let's see. The CIs of the plays in the [U V7] 20k rollout were (avg) .0055, and the CIs of the [U V15] 46k rollout are .004. That makes the CIs of the margins .0078 and .0056, and the CI of the difference in margins = sqrt(.0078^2 + .0056^2) = .0096.
That is, (assuming I've made no mistake), the probability of an independent 20k rollout and 46k rollout of the same position being .010 or more apart (as these are) is less than 5%, if you believe XG.
Anyway, after 67k trials, U (Up, Bar/21) beats V (Variant-Up, Bar/22 24/23) by .012.
[Btw, I hope that W (5k) is still part of your 46k rollout, and in order to emphasize the U/V margin you just chose not to display W this time; otherwise you might have two rollout files.]
Nack
| |
BGonline.org Forums is maintained by Stick with WebBBS 5.12.