| |
BGonline.org Forums
63S-43 started
Posted By: higonefive In Response To: 63S-43 started (neilkaz)
Date: Wednesday, 20 October 2010, at 7:18 p.m.
In the october supplement of the critical care medicine, i read in "Oiling the wheels of intensive care to reduce "machine friction": The best way to improve outcomes.": ...David Sackett (7) in an article discussing failed randomized controlled trials stated, “the confidence in the conclusion of a randomized controlled trial is the ratio of the magnitude of the signal to the magnitude of the noise times the square root of the sample size.” Here, “confidence” describes the confidence interval around the effect of the treatment in question, “signal” refers to the differences between the effects of the experimental and control treatments, and “noise” is the sum of all the factors (“sources of variation”) that affect a patient’s response to treatment. Therefore, extrapolating from this statement, the smaller the treatment effect, the larger the number of patient factors, and the smaller the trial size, the lower our confidence would be in its results. By the very nature of studies in intensive care, the “noise” part of the equation is very large. ..."
How can we sure, that this can also be a problem in openings RO. There is in my opinion a ton of "noise": the undefined state of the game, checkerplay handling of the bots, influence of live cube, evaluation settings, truncation settings, trials. If we have e developed position with key features, a rollout is in my opinion far more valuable. It is worth, to put so much energy in the opening? In the ICU, we work with bundles and we try to execute the basics in exellence. I am curious about the efforts in backgammon opening. Can you give me a short explanation for this efforts? Thank you.
| |
BGonline.org Forums is maintained by Stick with WebBBS 5.12.