This what I wanted to say: ---------------------------- walterjennings Curve fitting is good as long as the curve you are trying to fit to is a good representation of what you are trying to predict, as well as having a small enough grammar as to not be mapping directly to the sample set.
You may be confusing the term 'curve fitting', with the term 'over fitting'. Imagine I'm given a basket of fruit and I wish to predict whether they are apples or oranges, given our sample set, we notice that all the oranges are orange, so we make a rule in our prediction, if(color==orange) then fruit = orange else fruit = apple; which works perfectly with the sample set we are given, we would say the grammar used to describe our prediction is small, which is good, had we used an algorithm to analyse the color/prediction space, we would say we have been curve fitting to that solution space. So how good is this fit curve in the solution space of color to predicting apples vs oranges, well it turns out it is very good, even though it is possible to find an apple which is orange or an orange which is not orange, generally thats not the case. We find that even though we are highly fit to the curve we see in our solution space, it is still an extremely good predictor. Now imagine instead of using color, we use something like sample order as our predictive measure, so we find in our sample set, if we have two apples in a row, it is always followed by an orange. And we keep adding to our predictive grammar until we fully predict our sample set correctly. This would cause a much larger grammar, meaning it is less likely to be correct. We also see that this is not a good predictor in real life, and is highly 'over fit' to our sample set. The moral of the story, fitting to curves you see in your solution space can be an extremely good way of creating a predictor, as long as the curve you fit to represents something that actually has predictive value. Thus the difference between curve fitting and over fitting.
IMO that is not what the difference is. The two examples are disconnected. In the first example you are not making a prediction. You are identifying the property [apple] in terns of the property [color] given a rule that connects the two properties. In the second example you have a random sequence of colors and you are trying to predict the next color. If you are paid for every correct prediction and you pay for every bad one, then knowing that after two apples you get an orange with probability 1 and applying this you are fitting to the data. If the probability is not 1 but p, then the question is not whether fitting is good in general but whether the fitting you apply has enough predictive capacity to generate a positive expectancy. In this respect, I agree with you.
I would say fitting to the a color based solution space is still a prediction. Since it is possible to get orange apples and non orange oranges. So the probability of correct prediction in the real world is not 100%. The outliers are just fairly few.
How do you know you have an edge? Once you have the free time to go fishing for several days without worrying any ups and downs of markets. lol
You can say it is a prediction but its cause is inexact knowledge of properties, something that can be improved through research or by adding additional properties, like shape and shade of color. Still the algorithm to make the prediction will be very compact and this indicates we are not dealing with randomness or we are dealing with low degree of randomness. In the second case, you can have infinite many sequences that predict outcomes. In this case there is no algorithm possible to cover all cases and this is indication (but not proof) we are dealing with randomness. Regardless, curve-fitting is unavoidable in all cases. It is not an issue whether it is a bad practice but whether it has predictive power.
For a further opportunity for anyone to examine their belief system see: http://www.elitetrader.com/vb/showthread.php?s=&postid=3010251#post3010251
You have the edge when you were able to trade in a variety of market conditions; you are more than likely to have more winning day everyday than losing days.