Goomboo's Journal

[Goomboo]

Solution: don't trade.

Or if you trade trade very infrequently and more on long term cycles.

edit: You may want to include dates on your X-axis.

The frequency of trading is irrelevant for this example.  It's a comparison of two traders with the above-mentioned performance metrics after 100 trades.

The X-Axis has nothing to do with date - it's number of trades.  It's a random simulation in which the outcome of any given trade is random but the traders have structured their payouts as mentioned above.  It shows that there's a difference between being right and making money.

[Frozenlock]

And BTW, here you have the last three months with daily candles and 10 and 21 EMAs as per Goombo's strategy:



Pretty useless, isn't it?

It would have assured you a price of around $50-60, while the actual price could have simply crashed to the ground.
You can also see that he enjoyed a nice gain from the previous months.

[Goomboo]

Is it discipline to continue trading after 14 consecutive losing trades or is it a lack of self-awareness? Is the losing due to "randomness" or due to changing market conditions that erode system edge? Trading isn't like poker or gambling where the probability is fixed (for example: knowing that there are 52 cards, 13 of each suit, therefore you have X in the deck to make a flush, etc). The game can change the next day.

If you lose a coin flip by calling tails 100 times after 100 attempts, you're probably a sucker if you still think it's a fair coin.

Just something to think about. No right answers...

I would argue that it makes sense to continue trading under the following criteria:
- You have rigorously tested your edge
- You have risk control in place to trade less size during periods of drawdown
- Your method makes intuitive sense

I also would argue that it is simply due to randomness.  In my personal trading, I frequently will lose 7-8 trades in a row, but that's because I target a 1:5-20 risk/reward ratio.  My win percentage is 30-40% so it is perfectly normal for me to experience consecutive losers.

You're absolutely right: probability isn't fixed in trading - to the frustration of many participants.  All you really have is an educated guess based on history.

[Rampion]

And BTW, here you have the last three months with daily candles and 10 and 21 EMAs as per Goombo's strategy:



Pretty useless, isn't it?

It would have assured you a price of around $50-60, while the actual price could have simply crashed to the ground.
You can also see that he enjoyed a nice gain from the previous months.

Using the daily candles would have made you to sell at the very bottom ($50-60), I would have killed myself if I did such a mistake. Then, on the daily chart EMAs crossed again at $100-120, so you would have bought back at a huge loss.

To the contrary, using hourly candles you would have sold at $200ish (well.. If Gox queue let you, which is not obvious ), and bought back at $100ish after the bounce from $50.

I really do not think Goombo is using daily candles to check if the 10 and 21 EMAs crossed, can he confirm?

[nii236]

It looks like a crossover is imminent. Its been too long!

[Frozenlock]

It looks like a crossover is imminent. Its been too long!

Uh? There was an "imminent" crossover for almost 24h before the last little upward move.

[Psikopit]

Throughout my life, I have found that the true masters of fields are those who are able to express advanced concepts in terms that the common person can understand.  Trading is no different.  Without fail, the most successful traders I have met have been those who are able to take their decades of experience and logically present their method in less than 30 seconds to the listening ear.

The serious trader should ask himself if his method actually makes sense.

To my mind that's the key!!!! I'm a student (electronic PHD) and I have many teachers almost all are doctors (8 years or most studying a specific field). You can’t understand some of them even if they say they are detailing something very simple and easy but they take 1 hour to explain it. Others always start with “Look it’s very simple” and can detail something way more complex in 2 minutes.

In our case if your trading system is easy to understand then that don’t mean that he is to simple for trading but that this system make sense. That’s all. And with sense you’ll be able to keep trading even in a 7 loss row and keep your feelings apart.
If your system is too complex you’ll have much more chances to be confused and lost.

I’m just a trader for 15 days now so I can’t really speak with my experience but I’ve elaborated a bit the EMA system:
I’ve added a warning message (python bot) when I have BTC (that mean the EMA 10 is over the EMA 21) which calculate the differentiation of the EMA 10 (a*x+b function) if a<0 then the bot send me a message like “look at BTC price”.  This does not enter in the trading system I’ve just added it as a warning because I’ve noticed that under a certain a coefficient the market tend to be short quiet often.

With this method I can trade using a Hourly based EMA. And by the way I’m also considering to use 2 EMA base or more in the trading system. Like checking on a 4-hour base to have the trend of the market and selling when the 1-hour  base EMA are crossing. Is it an admissible idea?^^ I suppose you’ll tell me to back test this method. I tried to do so but failed when I needed to use Ninjatrader :/ is any one able to supply me some help or documentation on back testing using ninja or any other tool?

Eventually, on 10 days of active trading I made 3 full trades (USD->BTC->USD) and I had 6.7% profit, -4.3% loss, 7% Profit: ~9% of profit (including the Mt. Gox fees), I’m definitely a Pro trader^^ (just kidding). But I never though that trading would be following a system and keeping away “information influencing the price of BTC” and I support this now

[Goomboo]

With this method I can trade using a Hourly based EMA. And by the way I’m also considering to use 2 EMA base or more in the trading system. Like checking on a 4-hour base to have the trend of the market and selling when the 1-hour  base EMA are crossing. Is it an admissible idea?^^ I suppose you’ll tell me to back test this method. I tried to do so but failed when I needed to use Ninjatrader :/ is any one able to supply me some help or documentation on back testing using ninja or any other tool?

I'm happy to hear that you are profitably trading.

Your idea of looking at the trading system across different timeframes (1-hour, 4-hour) is certainly valid.  It's a concept known as "multiple timeframe analysis".  It's based on the theory that the trend of any given timeframe will more than likely be influenced by a higher timeframe.  In trading jargon, the trend of the higher timeframe is known as the "primary trend".  So, the idea would be to only hold BTC on the 1-hour timeframe if the 4-hour timeframe says you should hold as well.  Or in trader speak: "Only trade with the primary trend".

I do suggest testing it and you can use Excel, Ninjatrader, or even test it by hand.  You can express all of the moving averages in a single time period to make testing easier.  For example, the 10 period moving average on the 4-hour chart is the similar to the 40 period moving average on the 1-hour chart.

The general steps in NinjaTrader
-Import the data you're trying to test - it has to be in this format: http://www.ninjatrader.com/support/helpGuides/nt7/index.html?importing.htm
-Create the strategy using NinjaTrader's strategy builder (or code it yourself...strategy tester is pretty easy for most scripts...and NinjaTrader already has a moving average crossover pre-programmed)
-Backtest the strategy across the data you imported using the Strategy Analyzer

[Psikopit]

Ok for the multiple timeframe analysis and if something like this happen what should I do? (using this method)
http://bitcoincharts.com/charts/chart.png?width=940&m=mtgoxUSD&SubmitButton=Draw&r=5&i=6-hour&c=0&s=&e=&Prev=&Next=&t=S&b=&a1=EMA&m1=10&a2=EMA&m2=21&x=0&i1=&i2=&i3=&i4=&v=1&cv=0&ps=0&l=0&p=0&

http://bitcoincharts.com/charts/chart.png?width=940&m=mtgoxUSD&SubmitButton=Draw&r=2&i=Hourly&c=0&s=&e=&Prev=&Next=&t=S&b=&a1=EMA&m1=10&a2=EMA&m2=21&x=0&i1=&i2=&i3=&i4=&v=1&cv=0&ps=0&l=0&p=0&

higher time frame say "hold" and primary trend say "sell", I think it's sell.

edit:
I mean what information the longer time frame give? The trend of the market? That's what I suppose.

[Frozenlock]

You sell half? 

[Psikopit]

Buy half sell half in that case^^

[Goomboo]

Ok for the multiple timeframe analysis and if something like this happen what should I do? (using this method)

higher time frame say "hold" and primary trend say "sell", I think it's sell.

edit:
I mean what information the longer time frame give? The trend of the market? That's what I suppose.

Yeah, in that situation, you'd exit your position if you were relying on multiple timeframes.

The longer timeframe tells you which direction you should be trading.  The shorter timeframe tells you when you should be trading.

My personal opinion is that multiple timeframes tends to confuse things, so I prefer to stick with analyzing a single timeframe.  If you're able to use it to your advantage, then by all means, test and trade!

[Psikopit]

Ok for the multiple timeframe analysis and if something like this happen what should I do? (using this method)

higher time frame say "hold" and primary trend say "sell", I think it's sell.

edit:
I mean what information the longer time frame give? The trend of the market? That's what I suppose.

Yeah, in that situation, you'd exit your position if you were relying on multiple timeframes.

The longer timeframe tells you which direction you should be trading.  The shorter timeframe tells you when you should be trading.

My personal opinion is that multiple timeframes tends to confuse things, so I prefer to stick with analyzing a single timeframe.  If you're able to use it to your advantage, then by all means, test and trade!

Alright thank you for all this advises I'll test it and validate. It confuses a bit but if you're just looking at the direction of the highest one it's not that confusing
Thank you anyway for all this good tips you give and all this explications!

[pwi]

With this method I can trade using a Hourly based EMA. And by the way I’m also considering to use 2 EMA base or more in the trading system. Like checking on a 4-hour base to have the trend of the market and selling when the 1-hour  base EMA are crossing. Is it an admissible idea?^^ I suppose you’ll tell me to back test this method. I tried to do so but failed when I needed to use Ninjatrader :/ is any one able to supply me some help or documentation on back testing using ninja or any other tool?

I'm happy to hear that you are profitably trading.

Your idea of looking at the trading system across different timeframes (1-hour, 4-hour) is certainly valid.  It's a concept known as "multiple timeframe analysis".  It's based on the theory that the trend of any given timeframe will more than likely be influenced by a higher timeframe.  In trading jargon, the trend of the higher timeframe is known as the "primary trend".  So, the idea would be to only hold BTC on the 1-hour timeframe if the 4-hour timeframe says you should hold as well.  Or in trader speak: "Only trade with the primary trend".

I do suggest testing it and you can use Excel, Ninjatrader, or even test it by hand.  You can express all of the moving averages in a single time period to make testing easier.  For example, the 10 period moving average on the 4-hour chart is the similar to the 40 period moving average on the 1-hour chart.

The general steps in NinjaTrader
-Import the data you're trying to test - it has to be in this format: http://www.ninjatrader.com/support/helpGuides/nt7/index.html?importing.htm
-Create the strategy using NinjaTrader's strategy builder (or code it yourself...strategy tester is pretty easy for most scripts...and NinjaTrader already has a moving average crossover pre-programmed)
-Backtest the strategy across the data you imported using the Strategy Analyzer

Goomboo; you never crease to amaze me both quality of your analysis and your willingness to give it away for free. You make some awesome contributions that you not at all obligated to share.  Thank you.

The clear and concise way you present time as it is relevant to trading is the best I've seen this concept described.

In what language do you write analytic code to test current theory and past performance?. Is it the same as your predictive models?

In the rawest of forms, I have found Python to serve these purposes well 

Keep up the good work goomboo.

[stoto]

Hi all, Goomboo here!

I'm a Forex / Bitcoin trader and I thought I'd open a thread in which I share my method / trades.

I have tried trading thousands (literally, thousands) of trading strategies for the past 6 years, but I have found that the more simple and less elaborate I make my trading, the more profitable it becomes.  The system I use for Bitcoin is very, very simple and straight forward.

I trade a 10 and 21 exponential moving average crossover.  When the 10 crosses over the 21, I buy.  When the 21 crosses under the 10, I sell short.  I use hourly charts on www.bitcoincharts.com and guarantee that there is liquidity to prevent slippage on bitcoin.clarkmoody.com.


I have a small account in the Bitcoin market, which I hope to grow through trading!  If anyone makes money trading and they take a signal based on my recommendation, please consider donating a small amount of the profits!

Hey GoomBoo,

If you have some spare time, can you please take a quick look at http://bitcoin.stoto.net/ and give me some suggestion, what to add, to make it more useful for traders like you?

[Psikopit]

Draw EMA's? and the volume available at a certain order.

[Goomboo]

In what language do you write analytic code to test current theory and past performance?. Is it the same as your predictive models?

Glad you have enjoyed the thread, thank you for the compliments.

I test both in NinjaTrader and Microsoft Excel.  I use Excel for the majority of my testing / trading research.

Best of luck

[dragony]

Hi Goomboo,
Budhist says hello to Christian . Three past evenings I was reading this Journal. Every comment from every user and it was quite a ride I have to tell.

Red and green arrows where to buy/sell, EMA, beautiful colourful tables, predictions about BTC when it was much weaker and vision of it's crashing. Also few haters, many appreciations and kindness and calm of Gomboo.

I have to say thank you a lot for a great education. I am looking forward for some new knowledge:)

[ft73]

Hi Goomboo,
Budhist says hello to Christian . Three past evenings I was reading this Journal. Every comment from every user and it was quite a ride I have to tell.

Red and green arrows where to buy/sell, EMA, beautiful colourful tables, predictions about BTC when it was much weaker and vision of it's crashing. Also few haters, many appreciations and kindness and calm of Gomboo.

I have to say thank you a lot for a great education. I am looking forward for some new knowledge:)

Agree, best thread in the speculation subforum.

[ft73]

This thread helped me prototyping a very raw trading bot (for Bitstamp).
Here's its behaviour so far, while simulating 1 BTC trading over past 90 days.
It does miss a lot of the spiking, but over 100% gain isn't too bad, with few loss (4) vs gain (11) orders.

90 Days Bitstamp 1BTC trading Report (SIMULATION)

Now trying to improve it.