Since you are (I think) about to do the long awaited study, how about telling me in advance what different results will mean? I would hate to be in a no-win situation...
is this the same guy that was pushing to study \"standard deviation\" earlier? hopefully he has gotten some assistance in understanding statistics a little beyond a \"statistics for dummies\" book.
Jerry,
The only thing I am really interested in is finding out the relative standard deviations for TG and Rag. What different results will \'mean\' is completely open to interpretation.
Based on the vast majority of posts on this board and Ragozin\'s board, the expected result is that TG will have a significantly smaller standard deviation than Ragozin for each race type -- the \'smoother line\' effect. For my money, it is just an interesting fact to chase down -- just how big is the difference? If, in fact, the difference is not terribly significant, then that would be something of an \'AHA\'. However, there is nothing in the results that will help someone decide which product to buy, if that is what you are getting at. I started to input the numbers today. It will probably be another week before it\'s done.
As for you, Mr. Holbert, based on your posts seen this board, I would be very careful before engaging in a discussion of statistics. If you have a point to make, make it and back it up with something. Tossing out fact-poor -- or, in the case above, fact-less -- missives is just a big waste of cyberink and time.
Slow down there...
1- If the difference is not terribly significant, then that would be something of an aha to whom? Aha meaning what? What difference qualifies as \"terribly significant\".
2- If the difference IS terribly significant (as hopefully defined by you in advance), is that also an aha?
Jerry,
Maybe we could take bets:
What do you think the difference would be?
How about others on the TG board?
I can\'t think of any other way to determine whether the results will be surprising or enlightening to you or anyone else for that matter since we don\'t have any benchmarks. Suggestions?
Bottom line, though -- if this is not going to be interesting to anyone but me then I will cease and desist. This is a very very time consuming and detail-oriented piece of work.
David,
I\'ll go out on a limb. I don\'t have much interest in this. What are the possible outcomes and why would any of them influence a potential customer? If you proved out the \"smooth line\" effect, it would not demonstrate that either product was more or less valuable. I don\'t see how any possible result on standard deviation as I understand it will sway people one way or another.
I haven\'t posted in awhile, and I did notice something related to an earlier debate on jockey/ROI. I think we were in agreement on this (that it was not a terribly useful stat). Since it inspired such a passionate debate, I feel obliged to post this follow up.
I looked in the OTB program last week, and I noticed that John Velazquez is winning at about a 20% clip, same as he was at Saratoga. However, his ROI has plummeted to about $1.00 from about $2.00 in the summer. The jockeys that are leading in this category are named Castillo and Chavez (and that\'s Luis, not Jorge).
I think this shows that jockey/ROI is a reflection of factors that have nothing to do with the ability of the rider (which in Velazquez\' case remains constant - 20%). It IS a reflection of field size (Velazquez\' 2-1 Saratoga winners are now 3/5), competition (Velazquez doesn\'t stick out as much at Saratoga), the changing faces of the jockey colony (bye bye Bailey, Day and the various California heavies that make the summer scene) and the cyclical nature of the New York racing year (a combination of all of the above). This cycle is fairly predictable and Velazquez\' ROI will rise in the summer as surely as Luis Chavez\' will plummet. I won\'t call this stat entirely useless but it has to be ONE of the most useless stats in a sport that generates volumes of dubious data.
Hopefully this tangent wasn\'t too annoying. Pick some winners David. Have a Happy New Year. HP
you have a post in mind?
as has been asked by me and others, what do you think it means if one product has a larger standard deviation than the other?
here\'s the beauty in this kind of product:
let\'s say i have two people measure the lengths of all the pieces of straw in a bale. after they are done, i burn the straw so you have no way of knowing the true measure of any of the straw. now you take these people\'s measurements and put them in your newfound stddevp function in excel. what is this going to tell you? let\'s further state that the differences between the measurements are significant. what does this mean?
here\'s something that might have some interest and would be much easier: we can take the upcoming gulfstream meet and compile the percentage of winners with the best last race number. we\'ll toss in bris, the sheets, and thorograph. if someone wants to toss in beyer, that\'s fine also. if you only get the sheets on the weekends, we\'ll only use weekends. you want to use a different meet, we\'ll use a different meet.
this does nothing to address the pattern issue. i haven\'t thought about it enough to say it\'s valid or not. i just said there might be some interest here.
Only problem is I don\'t think any handicapping products really help at Gulfstream, especially at the beginning of the meet. HP
David-- just got back from a few days off. It doesn\'t make much difference to me whether you do the study or not, although I am curious about what it would show. But what I don\'t want to do is walk into a situation where one result makes Ragozin look good, and the other(s) means nothing. How about direct responses to the questions in my previous post? Since I don\'t know of any similar studies, I don\'t know myself what would be significant-- and if you feel the same way that\'s fine. But then there are no \"ahas\".
Jerry,
I thought that I implied an answer to your questions in my previous post but here are the direct answers:
You wrote:
1- If the difference is not terribly significant, then that would be something of an aha to whom? Aha meaning what? What difference qualifies as \"terribly significant\".
Per my last post, this is a subjective measurement, but, let\'s say that Ragozin and Thorograph lines were actually of similar \'smoothness\' e.g., std. dev. within 10% of each other. I think that would have qualified as an \'Aha,\' in that there was less difference between the two products than I believe most people would have thought. I can\'t prove this but I bet that if you asked people to take a stab at quantifying how much smoother TG lines are than Ragozin, most would say something significantly more than 10%.
2- If the difference IS terribly significant (as hopefully defined by you in advance), is that also an aha?
It might be, but if the prevailing view is that there is a big difference in std. dev. then the \'Aha\' would muted.
One last disclaimer before I toss the Sheets I printed out to do this study -- the point was never to make one product look good or bad. It was just to synthesize some of the differences in the results that the two methodologies produce, not from a betting results or numbers accuracy perspective.
I will be on betting hiatus for awhile, but look forward to cherry picking some posts that catch my fancy.
Actually, I don\'t really think you adressed my questions, or at least all of them, but it doesn\'t matter now. Too bad, because I was curious, \"ahas\" aside. I have no idea whether the methodology was correct or how much difference would be meaningful (we know that a 350 batting average is way better than 250, but what\'s way \"better\" here, even if one takes the position that something is better?), but it would have been interesting, and sparked discussion. Which would have been good, as long as no-one got sandbagged.