Which do you hate more?
Frankly I think the under optimizing is even worse...over-optimizing is a real concern too...oh well
Frankly I think the under optimizing is even worse...
interesting, pls explain a bit more by what you mean by "under-optimizing"
I guess, what I meant is: if you do not back test enough (at least 1 market cycle minimum), you are running the risk to constantly leave money on the table and limit your wining and consequently limit your ability to compound your winnings...
Quantt, would define each so we are all on same page.
over-optimization? under-optimization?
we all have our own concept of each, so lets say explicitly what we are each talking about.
One example would be betting size: if you read Alexander Elder's books where he is promoting risking only 1 or 2 percent, but this could be under optimization compared to using actual data and Kelly criterion for example to calculate the exact for betting size for your strategy...
If Kelly calculation is returning let's say 6% for back testing of 20 years - this could be an over optimization, because the future winning and loses cannot be quantified...
In this case as suggested as one of the first quants, maybe half a Kelly is the best to be used in the stock market - 3% in our example...
Have you read any of Nassim Taleb's books especially the one titled "Fooled by randomness"? Essentially in investing, it works until it doesn't and that randomness plays a much more significant role than is normally understood. What is a truly representative market cycle?
I still prefer trading on small amount of money, as it enables me to have a better control over the risks. If you risk everything, you can not only make a lot of money at once, but also loose it.