書中自有黃金屋 - Efficiency of Racetrack Betting Markets

[kido]

Using invesment market terms

(A) = technical analysis (think guys who look at price charts and talk "shoulders")
(B) = fundamental analysis (think guys who visit companies and read financial statements and create ratios)

It's hard to tell which approach is better because they're basically "climbing up the hill from 2 different side".  At first (B) sounds better because you're drilling into "why" a particular horse win, but since "luck" play a major role, and the number of examples to tune the model parameters are scarce, plus past performance (of particular horse) will change over time and doesn't imply future goodness, it would be difficult to arrive such models, or the model comes up with is not accurate.

(A) on the other hand, uses little domain knowledge.  It gets into think "how" I could profit from a particular strategy, the math model come up with is influenced lesser by change over time, (well, this needs to be proved, it's only my feeling) and the performance of a particular horse, and the environment.  Hence this model is easier to implement.

[chin]

In the general financial market, I don't know if the chartist used any holdout sample to test their forecasting techniques. A few people I know who use charts all just learn some patterns from books or other people, then eyeball the chart, and predict the whatever from that. In my opinion, this is just guessing and just as good/bad from the general public or sentiment. And just happened all these chartist I know are stock brokers, who really are not known to predict the market well.

And it just happened that I know a few fundamentists very well and they generally can consistently outperform the market. I think that if you invest knowing the true reason for the decision, you can sleep better without needing to look at the market everyday to check if your "pattern" hold.

For betting races, the closest to "technical" betting is arbitrage between pools.

[hangchoi]

In the general financial market, I don't know if the chartist used any holdout sample to test their forecasting techniques. A few people I know who use charts all just learn some patterns from books or other people, then eyeball the chart, and predict the whatever from that. In my opinion, this is just guessing and just as good/bad from the general public or sentiment. And just happened all these chartist I know are stock brokers, who really are not known to predict the market well.

And it just happened that I know a few fundamentists very well and they generally can consistently outperform the market. I think that if you invest knowing the true reason for the decision, you can sleep better without needing to look at the market everyday to check if your "pattern" hold.

For betting races, the closest to "technical" betting is arbitrage between pools.

Hmmm.....I don't think those chartists are doing a good job. I did read a few books about stock chart analysis and compared the "theory" with those analysis made by the chartists on TV. I can say that almost 90% of those "Chartist" are applying their "chart analysis" WRONGLY.........I don't know why....maybe they exercise some of their discretion on the chart theory......

An obvious example is that they use to say that the stock price will be up after a cup and handle pattern....but indeed there are at least 5 more criteria to meet for a cup and handle pattern to show a future uptrend.......

[kido]

In the general financial market, I don't know if the chartist used any holdout sample to test their forecasting techniques. A few people I know who use charts all just learn some patterns from books or other people, then eyeball the chart, and predict the whatever from that. In my opinion, this is just guessing and just as good/bad from the general public or sentiment. And just happened all these chartist I know are stock brokers, who really are not known to predict the market well.

And it just happened that I know a few fundamentists very well and they generally can consistently outperform the market. I think that if you invest knowing the true reason for the decision, you can sleep better without needing to look at the market everyday to check if your "pattern" hold.

For betting races, the closest to "technical" betting is arbitrage between pools.

Right, being a fundamentalist is at least more comfortable than a chartist because they know what they're doing.  Not just action simply by a spotting a certain 'pattern'.  In mention of the holdout sample, I could say there are certain methods to generate 'holdout' samples to test their model's validilities.  In comparing the races and the stock markets, I think stock market is far more complicated than a horse racing system.  You can say that the racing system is a small, closed system, factor affecting the system is limited. 

The only thing I usually agree with chartists are the 250-moving-average.  Using computer modeling, it's quite true that the bypassing the 250-day moving average is a good sign of turning market.

[wongyan]

pure chartist can really be a fatal fault.  Since all the "dealers" know all shapes of charts, they can easily create a situation for you to sell (or they called cut-loss) your stock at lower than reasonable price.  i.e. they can create fear when they want to be greedy.

[hangchoi]

pure chartist can really be a fatal fault.  Since all the "dealers" know all shapes of charts, they can easily create a situation for you to sell (or they called cut-loss) your stock at lower than reasonable price.  i.e. they can create fear when they want to be greedy.

Well....it is difficult to say....somehow they act the same way because they know the pattern. It is like chicken and egg problem.

[kido]

Going back to the book, I finally ordered this book from Amazon, hardcover, cost about HK$700 (including the shipping+handling), not as deal as previously mentioned.  This book should worthed to be owned, and I think this book is what I'm expected for, not previous 計得精彩.

[chin]

Going back to the book, I finally ordered this book from Amazon, hardcover, cost about HK$700 (including the shipping+handling), not as deal as previously mentioned.  This book should worthed to be owned, and I think this book is what I'm expected for, not previous 計得精彩.

陳文中毒！ 

[hangchoi]

  Luckily he was poisoned by the book, not by the horse racing........

[kido]

Luckily he was poisoned by the book, not by the horse racing........

Too early to say... This is a whole new subject to me.  Maybe I'm really poisoned by horse racing, because the ultimate fate of reading that book will lead to betting.