Quote from: jimmut on September 09, 2013, 10:37:45 AM

Quote from: seanrarey on September 08, 2013, 11:27:59 PM

Quote from: jimmut on September 08, 2013, 11:08:54 PM

You know a feature I think might be nice since all the exchanges work in tandem basically. Yesterday my BTC-e went to a buy signal but Bitstamp never did. And then BTC-E went back to sell today..So basically it was a loser trade. I would like a feature that my bot doesn't make a buy or sell without the other exchanges I'm trading on having the same signal. Does that make sense or do you think it could be helpful or not?

I did a lot of math today on what it would have taken to not take that BTC-e loss. In the end, I concluded that it was simply a false trend signal from unique activity on that exchange (possibly due to alt activity). There are going to be false trend signals with EMA... yep, statistics show 70% of the time.

People a lot smarter then I have tried to come up with "solutions" to help mitigate these, but I have never seen one that appreciably reduces the EMA error rate.

This ain't magic, it is just a mathematical formula that statistically will keep you ahead of the curve in the long run.

Dreary market last couple of days for sure.

That's what I was wondering if this was a unique occurrence or something that happened more often. And what you described makes sense and I guess it really won't come up to often. Like you said the market has been basically sideways with little movement. Is their a way a bot can tell the market isn't really moving fast in one direction or another and that maybe it should hold off on a buy or sell as it isn't really a trend? Or that it should do a smaller % of your funds till its sure of the trend and then do the rest. All most like an less aggressive setting? Like you said their have been smart people who explored these scenarios so I doubt it but still would like some ways to minimize the losses and maximize the profits.

The recommended starter settings are pretty conservative. With the .25 threshold & 2hr poll times, means at least 4hrs of confirmation of a trend before a trade... that is pretty cautious... but false and shallow trends will still occur. It is just part of the game.