Ernie: No time to reply now; a shipload of errands to run . But what you have said is so "right on" that comment is in order. Will try to say something meaningful later. Billy ----------------------------------------------------------------------- In a message dated 4/7/2010 2:44:38 P.M. Pacific Daylight Time, [email protected] writes:
HI Billy, On Apr 7, 2010, at 2:21 PM, [email protected]_ (mailto:[email protected]) wrote: One suggestion that was never acted upon, or no better than a pilot study, was for creation of a forecasters "batting average" service to track forecasts and list %s for all name forecasters, year by year. I think, but cannot say for sure, that this never got far because of opposition from entrenched futures think tanks who, if wrong didn't want the world to know, therefore could be free to trumpet their correct forecasts as if that was the whole story. Yeah, academics are like that. :-P Maybe some other reasons this excellent idea never got anywhere, but the fact is that it didn't. But that is what we all need. Obviously it would be a huge help in the financial sector. Or does anyone know of some kind of tracking service which covers finance ? I mean a "good" service, not just one that looks at selected results -which, clearly, can be skewed any which way. A batting average service would amplify weak signals into strong ones that could not be ignored. What is shocking is that the *professionals* are like that too -- and not just the producers, but the consumers. I actually talked to someone who worked at one of those rating services, to try to understand their business model. Those folks who issue "buy" and "sell" ratings only get paid if someone actually acts on the information, by getting the commission (or a slice thereof) from the transaction. Now, since we're talking billions upon billions of dollars, and specific quantitive predictions that would be trivial to track, you'd think Wall Street would have a robust model tracking the performance of specific analysts. No. Nothing. Nada. Zip. As far as I can tell, nobody even *wants* to get quantitative measures of how well analysts do; they'd prefer to make decisions based on their gut. Even if that meant over-valuing short-term results. That to me was dumbfounding. It felt like Moneyball all over again, where the old hands insisted on making "instinct-based" decisions, rather than applying modern statistic tools. Now, maybe I just lucked into a backwards corner of the finance industry for my informant; but I don't think so. As sophisticated as Wall Street is when it comes to analyzing *money*, I suspect they are as bad as the rest of us when it comes to analyzing *themselves*. Even when there are likely *huge* financial incentives for doing so. I hate to say it, but I fear that Wall Street's loathing of self-analysis may well be a deeper problem than corruption or even economic short-sightedness. Worse, it is not just "them"; it is endemic to American culture, and arguably even human nature. The only think *I* know that can combat this is a culture of peer-review and professionalism, which can only be injected by a successful upstart competitor. Someone needs to do for financial forecasting what _538.com_ (http://538.com/) did for political polling, and then make a ton of money at it so that everyone else sits up and takes notice. Paging Nate Silver... -- Ernie P. -- Ernie P. = _______________________________________________ Centroids mailing list: [email protected] http://radicalcentrism.com/mailman/listinfo/centroids_radicalcentrism.com Archives at http://radicalcentrism.org/pipermail/centroids_radicalcentrism.com/
_______________________________________________ Centroids mailing list: [email protected] http://radicalcentrism.com/mailman/listinfo/centroids_radicalcentrism.com Archives at http://radicalcentrism.org/pipermail/centroids_radicalcentrism.com/
