eally there and not just some aberration.

Neural nets are not the "holy grail." Some think they will solve all they're trading problems. They won't. But they do offer some value.

You see, you can load up a lot of data i

A neural net is like a brain. At least that is supposed to be what it is like. It tries to make lateral connections in data. So in that regard, it is trying to simulate what a brain would do.

The thing is, a computer is stupid. I can't really think laterally. And that is the problem with a neural net. Let me give you an example to make it clearer.

Let's say that you feed data about price, the RSI indicator, some moving averages, and some custom indicator value into a neural net. What it's going to do is "crunch" those numbers. It's going to see if it can use the data you gave it to predict what will happen next with any level of accuracy. Done properly, it will find a connection in the data, if any connection exists.

If none exists, it will get confused, and be unable to "learn" how to predict. If it does find the connection, you then need to feed it new data to make sure that the connection is really there and not just some aberration.

Neural nets are not the "holy grail." Some think they will solve all they're trading problems. They won't. But they do offer some value.

You see, you can load up a lot of data in

d. I can't really think laterally. And that is the problem with a neural net. Let me give you an example to make it clearer.

Let's say that you feed data about price, the RSI indicator, some moving averages, and some custom indicator value into a neural net. What it's going to do is "crunch" those numbers. It's going to see if it can use the data you gave it to predict what will happen next with any level of accuracy. Done properly, it will find a connection in the data, if any connection exists.

If none exists, it will get confused, and be unable to "learn" how to predict. If it does find the connection, you then need to feed it new data to make sure that the connection is really there and not just some aberration.

Neural nets are not the "holy grail." Some think they will solve all they're trading problems. They won't. But they do offer some value.

You see, you can load up a lot of data i

dicator value into a neural net. What it's going to do is "crunch" those numbers. It's going to see if it can use the data you gave it to predict what will happen next with any level of accuracy. Done properly, it will find a connection in the data, if any connection exists.

If none exists, it will get confused, and be unable to "learn" how to predict. If it does find the connection, you then need to feed it new data to make sure that the connection is really there and not just some aberration.

Neural nets are not the "holy grail." Some think they will solve all they're trading problems. They won't. But they do offer some value.

You see, you can load up a lot of data i

tion in the data, if any connection exists.

If none exists, it will get confused, and be unable to "learn" how to predict. If it does find the connection, you then need to feed it new data to make sure that the connection is really there and not just some aberration.

Neural nets are not the "holy grail." Some think they will solve all they're trading problems. They won't. But they do offer some value.

You see, you can load up a lot of data i

eally there and not just some aberration.

Neural nets are not the "holy grail." Some think they will solve all they're trading problems. They won't. But they do offer some value.

You see, you can load up a lot of data into a net. All kinds of things you would never think about. Anything you assume to be unrelated to the market (who won the Superbowl, what phase the moon is in, etc), and you can let the computer work away at to conclusively say one way or another if there is really a connection to the market's price or not.