Topic: Goomboo's Journal - page 16. (Read 281467 times)

I understand and I'll back off - I hope I didn't ruffle any feathers, I just know how people think. They don't do their research: greed takes over and they act first and learn their lessons later. Just like in other parts of the Bitcoin world - it makes me sick seeing people lose tens of thousands of dollars from their Bitcoin Wallets because they didn't appreciate security sufficiently. In addition there are those people that buy FPGA and even Video Card based systems for big money on eBay without spending 15 minutes researching what type of returns they'll get and how long the payoff will be (or whether it actually make more compared with power costs).

I did what others would do which was to read the basics on the first page and go to the last page and see the charts, get excited and be ready to trade. I can guarantee you there are numerous people that are trading it right now without doing analysis and without REALLY understanding what they are trading and what the risks are. That's why I asked the questions - they are tough questions, but in reality, I believe they are THE questions everyone should be asking. They need to say "that's awesome, but look at how 80% of the profit is lost to commissions - what type of risk am I looking at?" and need to understand the disclaimers that you accurately show. They need to appreciate it all before allowing themselves to trade a single Bitcoin.

Look - I'm certainly no big time trader, but I AM experienced. In 2008, I programmed my own systems to store, analyze and trade the Forex market. I had some very creative and very impressive (if I say so myself) pieces of technology, but it didn't matter. I didn't appreciate what and how I was trading and my system backfired and I lost massive amounts of money. I was very fortunate as I was able to fully recover and it didn't cost me my house, my family nor my future - many others are NOT so lucky and lose everything. That may seem like an exaggeration, but greed is a powerful emotion and the idea of having 500% returns in a short period of time is tough to pass up.

As I said, I'll back off now. Please be careful, cautious and don't let greed take over!

Two things: (1) I missed the part where Goomboo switched from 1h to 1d. Big difference already Glad Goomboo cleared that up for me. (2) If you read my post, you will notice I wasn't strictly speaking about total profitability alone, but *total number of trades* and *share* of profitable trades. The latter two are important mainly for psychological reasons, for me. Thats the point I was trying to make.

Oda.krell you might be wrong 5 times now - this is lets say 10% percentage of loss. But if you listen to Goomboo being wrong doesn't matter because you would earn 40% and another 10% in the meantime (real figures in july). So what is the point of being wrong? Read again previous Goomboo's post about fast loss cutting and long rally profiting.

I certainly understand - not every method works for every market participant.  What works for me absolutely won't work for all.  That's the great thing about markets - we can all have varying methods and still profit in the long run.  You are right to be critical of it and I hope that you find a method that works well with your style.

My personality is to follow the trend with clear-cut rules and this method works well for me for this reason.  If my personality was different, I simply could not trade this method and would have to find another.

For the reasons you mentioned (too many trades both positive and negative), I trade on the daily timeframe.  I don't trade hourly.  The thread started as an hourly method but a few weeks after the initial post I transitioned over to the daily timeframe.  The hourly simply doesn't suit my preferences - too many trades, too much screen time.

The key traits - discipline, consistency, confidence - are universal across styles, but the application of these traits will vary from person to person.  The world would be boring otherwise

Best of luck!

First time I'm posting in this thread, but I've been checking it for a while. I appreciate what you do here a lot -- first, rigorously introducing the EMA20+10 average crossover method, which is a staple technique, secondly, maybe even more important to me, stressing how much testing, questioning your own methods, and discipline in executing your method matters. Thanks for all of this.

Hope you don't mind that the following will be somewhat critical of the method you mention throughout this thread, the EMA20+10 crossover method. Here's the short version of my problem with it: it gives too many false positives. And, applied to a useful time interval (1h), it is awkward to use manually.

Here's the longer version  ... just happened again these days. Using 1h as the time interval, the EMA method interpreted the recent dip as a trend to trade on. Selling crossover was at pretty much exactly 106, buying back at 106. Including slippage and fees, that's a (small) loss, almost certainly. This happens quite a lot with this method, in my opinion.

Here's the second complaint: using the method on 1h time interval,  because of the high sensitivity of the method, you either have to constantly watch the price or be able to trade, or you only check it, say, once a day, but then you're lagging behind.

In conclusion, I started my btc trading journey learning about the EMA20+10 crossover method, but I've given up on it pretty soon afterwards. Maybe it's just my personal preference, but while I still work with moving averages, I prefer slower moving, less sensitive parameters. They net less of a profit than the EMA20+10 method at it's best (i.e. I sell and buy a bit too late, compared to the faster EMA), but my share of profitable trades is higher, I believe, and my total number of trades is lower, which stresses me out less.

The point of this thread isn't to say that a specific combination of indicators is profitable - it's to teach you what needs to be done so that you can become profitable.  Please note the subtly of that previous sentence.  Successful trading isn't found in a particular method.  It's found in the consistency, confidence, and discipline to backtest and act on what you know to be true.

I'll use myself as an example - I knew that moving averages backtested profitably years before I actually started earning a profit in my account.  I became a successful trader when I learned how to persevere during periods of difficulty.  Over the past 3 months, I have only won 22% of all trades which I have attempted.  In fact, I am currently in the midst of an 11 trade losing streak - I have lost the last 11 trades I have attempted.  When I was a young trader, these difficult markets devastated me.  I was unable to continue executing the next trade because I couldn't take the pain of being wrong and losing money again and again and again.  However, through time I have learned to focus on acting out what I know to be true and letting the results take care of themselves. 

It is very important to note that since I have confidence in my method, I took two trades in the midst of a string of losses that I am still holding which are propelling my account towards new highs.  So at the end of this quarter, I'll probably have only won 20-30% of my trades but ended up profitable due to the determination to act out what I have tested extensively.  The secret?  I cut my losses and let my profits ride.  Each of my losers were cut within hours or minutes while my winners have been held for several weeks.  It really is just a numbers game, but you have to have discipline to play it well.

Thanks for the question,

Your math is a bit off.  You've done some studying on the typical bid-ask spread, so let's use your figure of .28%.  In essence, this is an added commission in that you will experience an added loss of .28% on each side of the transaction.  Here is the updated chart showing a .60% commission and .28% bid-ask spread:




It certainly does hamper performance, however hourly traders are still up 210% year-to-date after factoring in your slippage figure.  This really isn't the point of the thread though.

I highly encourage you to read over this thread.  The point of this thread isn't the 10/21 hourly crossover.  In fact, I demonstrated a few months ago that there are better combinations than the 10/21.  For example, if you had traded a 20/50 crossover (arbitrary choice) with this same commission and .28% slippage, you would have earned an 815% return over the same time period.  The point of this thread isn't a specific combination and timeframe - the point is to learn the tools and techniques to be self-sufficient.

This said, I feel that you are intensely focused on the "10/21 hourly crossover".  I laud your desire to learn and successfully trade but I believe that you are better served by stepping back and trying to understand the concepts conveyed in this thread.  For example, I believe that you are better off knowing how to backtest your own ideas across any timeframe rather than knowing that at one time the 10/21 crossover was profitable in the BTC market.

Also, you note that the 10/21 may not be profitable in the future.  Again, I have reiterated this several times and pointed out that the 10/21 hourly crossover is not profitable in major instruments like the EUR/USD.  The reason that the daily crossover has continued to generate profits for decades on end in the majority of markets gives me confidence to continue trading it in BTC.

---

Also, I notice that your account seems to have been made exclusively to talk to me:


If anyone else finds themselves in a similar situation, I encourage you to PM me directly or send me an email.  I'm not so important that the typical new member request process should be circumvented.

Ah, Excellent! Thanks for getting back to me. Maybe if you can just add a reasonable bid-ask loss, we'll be set then.

Certainly, there's no doubt that Moving Averages work and can definitely proven in the Forex arena. My concern, however, is the high costs involved. Look at your most recent results:

Gross Revenue: 2,437%
Net Revenue: 484% (minus bid/ask)
Commission Loss: 1,953%

What that means to me (correct me if I'm wrong) is that if instead of 2,437% you "only" earned 1800% you would've been looking at:

Gross Revenue: 1,800%
Commission Loss: 1,953%
Net Revenue: -153%

Meaning that you're done: 0 BTC - game over.

This also concerns me because the data involves a period where Bitcoin was gaining public knowledge and gaining rapid value increase and big swings likely resulting in larger gains. Now that Bitcoin is (hopefully) heading towards a more stable price, I'm wondering (out loud) if that means that larger net losses are likely once its tougher to get larger swings.

Again: thinking out loud, please correct me

That's simply not true.

I was an investor and can attest that the returns you see on a trader are correct - these are returns that they earn.

They are probably referencing the glowing advertisement on the front page "here's how our trade leaders have done against CTAs, gold, S&P 500".  That figure is a straight average of leader performance and doens't account for leaders who have been removed.  This is where the reviews probably come from.  I'd straight up ignore those figures (since it's window dressing/hindsight bias) and focus instead on individual trade leaders if you are actually interested in the company.

Anyway, I'm not a spokesman for the firm so that's all I'll say.  Also, this isn't the point of the thread, so I'm not going to argue the merits of the firm any further - let the buyer beware.  I'm not a financial advisor, do your own due diligence.

Best of luck

I will try to get to this post when I have some free time, but for now, here's an executive summary:
- The results are long only (no shorting included)
- The results account for a commission on both sides (no adjustment necessary)
- I highly encourage everyone to be very skeptical of results on lower timeframes - if you have a delayed execution by even a second, the results will be materially different.  This is why I prefer the higher timeframes and trade on daily bars - commissions, slippage, delays in execution, etc. take a much smaller chunk of your overall return.

But reading the reviews from the link above seems to indicate that the returns are not "real", but instead are manipulated after the fact.

You don't actually give Currensee your money - you just open a brokerage account and sign a limited power of attorney document allowing people who you select to trade your funds.  Also, the returns and fees are real.

But in general, you're right - most are better off parking their funds in an index fund with a low expense ratio and looking at it in 30 years.  But just like most complex things in life: "it depends".

Greetings and thank you for the feedback,

That post was over a year and a half ago, so I have updated feedback.  I change my recommendation from "highly highly" to "be wise".

I was an investor for a time with Currensee.  The issue with Currensee isn't the business model (the model is brilliant), the issue lies with the Trade Leaders and how people select leaders to manage their money.

The returns you see on the website are real (they use mark-to-market reporting so you can't hide bad trades), but the model tends to give incentive to trader who makes several trades per day and wins 70% or more of their trades.  It isn't that hard to win 70% of your trades if you win $10 when you're right but lose $200 when you're wrong :p

This model results in a vicious cycle of Trade Leaders - a person has a good 6 month track record due to luck, he is placed on the leader board, he blows up because he doesn't understand risk control and isn't able to cut his losses.  Currensee is catching onto this fact and they have started being more selective in their process.  It is my belief that these are growth pains of the company and eventually they will learn to select leaders better.

In my opinion, the way to avoid this cycle is to only invest with leaders who have been on the list longer than a year and have max winning trades significantly larger than max losing trades.

I closed my investor account because I believe myself to be a better risk/return proposition.  If I would have stuck with my initial trade leader (the most popular guy) and walked away, I would be up 135% (before fees) to-date.

The reason I was so enthusiastic about the company in the past (and I still am), is that it reward for performance only.  But as always, "buyer beware!"

I can't speak for Currensee nor for Goomboo, but I would doubt there are ANY legitimate companies out there where you can give them money and not get ripped off. This is true from the false claims (gotta love that ad I see for some guy claiming 99% "Profitable" returns) all the way up to major Mutual Fund providers that can't even keep up with the major indexes. Companies that do make legitimate profit will continue to increase/structure fees so that you get less and less. I believe that this is one reason why Goomboo started the thread, showing tests that you can run yourself and design your own system.

PS - Anyone have any comments on my questions above? I'm very eager to hear responses.

Goomboo,

Thanks for all the useful advice in this thread.
Since I have some money to invest, I did some investigation into Currensee. I found some reviews that were a bit shocking: http://www.forexpeacearmy.com/public/review/www.Currensee.com
Can you perhaps talk a bit about your experience with Currensee? Are you an investor or a trade leader? Thanks for your answer.

Goomboo, thanks so much for the table of contents!!


I like you right now.

First off, thank you so much to Goomboo for providing this strategy - I'm sure it is appreciated by all!

I am a first time poster (just got off "newbie" status), but found your post and was very excited by the chart and it got me started to create my own data gatherer from MtGox with the intention of eventually creating an automated bot.

However, as I got further into thinking about it, important questions got raised. Could you (and others reading this) please address my thoughts?

I'd like this to be a PRACTICAL analysis as opposed to theory because if real people are putting real money into this system, they need to be prepared for actual statistical results. Users who starting using your system prior to the big bubble certainly did well, but now the question is how will this perform ongoing and for people just now starting.

#1 - Your latest graph (page 43) shows results the 0.6% Commission and admits that there is no slippage from bid-ask spread. I do not believe this is realistic. If you're going to have a trading system, unless there is a clear value that you can place a Limit order for, it'll be a Market order and therefore will incur a Bid/Ask Loss.

I've only been storing one minute data for a day now, but so far after 1,191 points that looks to be 0.28% loss.

#2 - You assume 0.6% Commission, but doesn't MtGox take Commission from both Buys and Sells? For instance, lets say you buy 1 bTC on Day 1 and sell it 12 hours later. Wouldn't they tack on 0.6% on the buy and 0.6% on the sell? My tests buying and then immediately selling 0.01 and 0.1 BTC seems to confirm that. My typical loss was over 1.2%.

I believe, therefore, you need to assume a 1.2% Commission

#3 - I understand that you can "short" using 3rd party companies, but unless there is a way to do it directly in MtGox (or whatever your model is based on) I don't think you should include shorting profit. That's not much of a concern up until recently with the big bubble, but going forward shorting profits might be higher than long profits.

#4 - That's neat that you calculated it for the 0.25% Fee schedule, but lets face it - that's not realistic. 500,000 BTC per month - that's $54 million! Not only is that an absurd amount of money, but I doubt you could trade anywhere close to that without huge slippage.

So, if I'm correct, then that means you're really looking at a 1.5% commission rate or so. Could you re-calculate it for that commission rate? Unfortunately, I'd bet that'll make the numbers pretty weak, especially starting in May where things have calmed down. If you were to average one trade per day, 30 trades per month, that'd mean you'd need a 45% gross return every month just to break even! That's a minimum of 548% return required per year.

PLEASE tell me I'm wrong (and most importantly why) as I'd love to start making some good green, but unfortunately I think this is the way it works.

I have lots of strategies in Forex where it'd gross me a 10,000% profit per year, but when you include commissions and bid/ask loss it winds up being a 50,000% LOSS. Once BTC gets to the point to where it can be traded efficently, the big gaps in the exchanges (ie: the $8 difference between MtGox and BitStamp and the $10-$11 difference between MtGox and BTCe) will be eliminated and systematic traders will come out of the woodwork to capitalize on these trends which will eliminate the profitability of them.

PS - Could you show some monthly status from 2011? That might give us a good idea of what it looked like before the big bubble.

Thanks!

Comments?

I see what you're saying - my phrasing was a bit confusing.  I've changed the wording - thanks for the heads up!

Thank you for the feedback, I am glad to hear that you find the thread helpful.

I have added this list of topics to the first post in the thread and I will periodically update it from there.