Topic: Goomboo's Journal - page 46. (Read 281470 times)

Thanks for the question Clipse,

This is known as scaling in.  Basically it's adding to or building into a position.  The opposite of it is called scaling out - cutting your position as time unfolds.  Just like everything else, I'd recommend that you systematically test the concept and be sure that you are accounting for risk at every step.  After all, if you scale into a long position, you are increasing the average entry price of your trade, meaning that you could turn a winning trade into a loser.

For example, if you buy 100 BTC at $1, your average price is $1.00 per BTC.  The price goes up to $2 and you buy 100 more BTC.  Your average price is now $1.50.  Let's say price reverses and your exit signal occurs at $1.25 - you lost on this trade.

Here's some general and specific recommendations.  Let's start with specific.  Prior to entering a trade, if you are planning on scaling in, you should select a criteria which designates location as well as the amount you will scale in.  A very specific criteria I suggest is using the Average True Range (ATR) indicator to determine locations to scale in.  I use multiples of it at the time I enter a trade to determine scale in locations.  I use the 14 period ATR for this.  Period doesn't particularly matter, it should just make sense.

Example trade and scale-in:
-You enter the trade long at $3.06 and the ATR at time of the trade is $.26
-Scale in Location 1: Close + 2 * ATR = $3.58
-Scale in Location 2: Close + 4 * ATR = $4.10
-Scale in Location 3: Close + 6 * ATR = $4.62
-Scale in Location 4: Close + 8 * ATR = $5.14
--When price reaches these locations and you have an existing trade, you will buy / short another trade unit.  I suggest that this unit be equivalent or less than your initial size in the trade.
--For short trades, subtract the ATR value

General recommendation - every time price clears a prior retracement in the direction of the trend, then you can add to the position.  This is very subjective however and more vulnerable to emotional decision making.

I suggest not scaling more than 4 times.  The idea of scaling is that if the market strongly moves in a direction, you will be in it with progressively more capital.  If the market doesn't move very strongly, however, you will find that you have a higher percentage of losing trades.

exactly silverstew

that is the tradeoff; more chop and earlier trades, or less chop and missing more of the move

it all comes down to your appetite/tolerance for action

I noticed your a bit of a cross fan.

Seems a little choppy to use the shorter numbers, but might help validate earlier entry points?

i have a basic and probably dumb question. this exponential EMA average, how are the weights? how is this decay function defined and does it also have parameters additionally to the time-window?

I agree there is nothing magical however its a measure and I figured to add more to my position there could be something to look at without doing it blindly. I will have a look at your suggested 15-30 cross, thanks.

there is nothing magical about the 10-21 cross

generally shorter averages will get you in or out earlier, but will have more zig zag trades

with longer averages you will miss more of the beginning of a trend, but there is less noise so more of your trades will be keepers

so for the above poster, maybe look for a 15-30 cross or something to confirm for adding to your position

Goomboo do you have any suggestion on using the EMA10/EMA21 as described in this thread after it allready cross up/down to predict longer rally periods ie. I bought 50BTC as the EMA10 cross upwards over EMA21 however I then notice the line keeps trending upwards but now I want to buy more BTC to take advantage of longer rally however I dont know if at the time I buy more BTC the EMA10 will cross down over EMA21. Do you know of any strategy to make use of longer rally cross so I can add more BTC to my long(or short in case of downtrend) even after the cross happened.

yes, that depends on the publisher and the country where you are. e.g. i see qutte a few pages over here, but a lot of them are just missing from the "preview".

I'm a trend follower, in this mornings case the extra indicators were added confidence but wouldn't necessarily dictated my action. Thanks for this thread by the way, it's good to have an extremely experienced person around to recommend the right books. I only wish I had the discipline to follow your approach and back test my strategy which I haven't even defined yet. But then again, that's the beauty of bitcoin, with only .65% commission at Gox, I have thousands of trades experience which I wouldn't have anywhere else along with the fact I can jump in and out. Were I paying 10$ a trade, along with account fee's I suspect I would be far more strict in my approach.

Sometimes whole books are on google, other times huge previews.

Hmm, this is odd - I was searching for a study he once did on random entries and I came across a section of the book on Google...is this legal?  Can you get a whole book on Google these days?  If you can, I strongly suggest purchasing the book - supporting the author is where it's at.

http://books.google.com/books?id=dTmr94EhEoUC&pg=PA187&lpg=PA187&dq=thomas+stridsman+random+entries&source=bl&ots=bwcHqGL-h8&sig=iKT7QHIhaxAoHVmkXJT04PFvm18&hl=en&sa=X&ei=zCQkT9Qd84-wAqebqYwC&ved=0CCAQ6AEwAA#v=onepage&q&f=false

Yeah, indicators often tell the same story.  After all they mostly share the same purpose - trend following or overbought / oversold. 

I'm wary of looking at too many indicators, personally.  An author of an excellent book made a good point when he talked about indicators.  He basically said that if you have 5 indicators, each with a 30% chance of calling the trend, and you rely equally on all of them, you have a less than 1% chance of actually calling a move.

If anyone is interested in some graduate level system design, this book is it:
 
Trading Systems that Work
http://www.amazon.com/Trading-Systems-That-Work-Evaluating/dp/007135980X

Yeah, that's lame :/

Thank god for the crossover alarm, twas in the middle of the night here, MST.

just before this someone put up a 20k or so bid wall around $5.30 I think it was, right after my little buy it disappeared & the spike volume buy happened, same person I guess



 

Yeah, that's an interesting one, shortly after someone went 20K BTC in leading to the spike.

Thing is.....There's more than the 10/21 EMA crossover here. Aroon UP/Down also crossed over and the MACD also hit the positive line around the same time so who ever played the 20K might have been working on a different strategy.

The price is still outside the Bollinger band to the positive, so with this many positive indicators it will be an interesting weekend for the short term.

My bets: Stop Loss 5.45 - Arbitrary based on my entry point. 6 will likely be tested.

EMA(10) > EMA(21)  



mine was the little buy that pushed the price from $5.40 to $5.45 (821 BTC took just under 11 seconds) & I'd like to say a big thank you to whoever sailed in right afterwards with the 23k spike it to $5.75

Well, this is the first time I've ever actually imported BTC into NinjaTrader - since I'm on daily candles and a normal signal lasts about a month, I'd probably just repeat the steps and re-download etc. when a month has past or so.

Too bad NinjaTrader is Windows only.  I'd like to try it out.

NinjaTrader is great for non-programmers as well.  You can build an entire strategy from point-and-click as well as perform backtesting - basically a platform made for traders by traders.

R sounds cool too - lots of tools available for people to do analysis.