Step 1.First, you need a charting service. Since most all electronic trading platforms have charts with technical indicators, this shouldn't be a problem.

Create a 1 hour chart on whatever currency pairs interest you. Barcharts or candlesticks really make no difference. Overlay on this 3 things: 1) a 169 period [1 hour] ema [exponential moving average], 2) a 144 period [1 hour] ema, and finally 3) a 12 period  [1hour] ema.

The 144 and 169 ema's create what I call the "tunnel". The 12 ema is an extremely valuable filter that you will want to have there all the time. I will talk more about this in the filter section.

Step 2.Memorize or write down and keep next to your trading screen the following fibonacci number sequence: 1,1,2,3,5,8,13,21,34,55,89,144,233,377. For trading purposes, the numbers of  interest are 55, 89, 144, 233, and 377. 【BINGO 收集整理】

Step 3.Wait for the market to come into the area of the "tunnel". When it breaks ABOVE the upper tunnel boundary, you go long. When it breaks BELOW the lower tunnel boundary, you go short.

Step 4.Stops and reverse are placed on the other side of the tunnel. 

Step 5.As the market trades in your direction, you take partial profits at the successive fib numbers respectively, with the final portion of your position left on until one of the following conditons occur: 1) market hits the last fib number [377 pips]  from the ema's, or 2) the market eventually comes back to the tunnel and violates the other side.

Example: GBP/USD is trading at 1.8500. The ema's are as follows: 144- 1.8494, 169- 1.8512. The market breaks 1.8494, and you sell at 1.8492. Your stop and reverse is now at 1.8512. Over the following hours, market starts to go down. 40 minutes after you put position on, cable is at 1.8440. You can use for computation purposes either tunnel boundary or the median of the tunnel. Ema's are still the same, so if you use the median, 55 from 1.8503 is 1.8448. You should have taken part of the position off at 1.8448. Market does nothing rest of day. Stop can be moved down to protect position or left alone at tunnel.  You are now looking for price to be 89 pips away from the ema's. Since 55 was already passed, it no longer concerns us in this cycle. A couple of days later, cable is at 1.8300 and the median of ema's is 1.8410 [1.8400 - 1.8420]. You should be out of another portion of the position at 1.8321. Market bottoms here and in the next 2 hours, cable screams to 1.8535. Your remaining short position is covered at upper tunnel boundary of 1.8420, and you are now long from this point as well. Since you are long, you would now take partial profits at 1.8475 and 1.8509.

This is a fairly typical example.

If you were to just stick to this basic model, you account would grow very well over time.  Las Vegas was built with far fewer percentages in the casino's  favor.

In case you haven't figured it out, this model cuts your losses very short. By definiton, you can't lose very much on a single trade from your initial entry position.. On the other side, you take some quick profits at the 55 level which satisfies the scalper in you, and you have positioned yourself for bigger profits in the long run should the market keep going in your favor. By definition, you are letting profits run.

The Achilles heal of this model is when the market chops around the tunnel and gets you in and out multiple times for small losses. I will cover how to deal with this in the filters section.

That's it. This is the model. Fairly simple in its design, and easy to remember. Has all the things every local wants in a model, except the quick 2 pip scalps, which you can't do anyway. Cuts losses, and lets profits run. Yet for its design simplicity, the thought behind is more complex. Time to talk about that.

THE FILTERS

Filters are used to increase overall profitability and/or reduce overall losses. If a filter does not do one of these two things, then I do not use it. What good is a filter if it raises your profitability by 10% but only gets you into 1/3 as many trades? What good is a filter if it reduces losses by 10% - 20% , but also reduces profitability on every trade by half? I think you get the point.

Here are the filters the vegas team uses. [Yes, I have a team. There are 3 of us. We trade GBP/USD, USD/CHF, and the S&P e-mini futures contract. Each has a specialty. Mine is GBP/USD. We are each responsible for our main pair. One of us is always at the screen when markets are open. Positions are covered by other partners when away. We only tunnel trade.]

Put the 12 ema [1 hour] on your screen with the rest of your indicators. When everything is at the same price [tunnel, current market price, 12 ema] sit up and take notice. When the market breaks away from the tunnel, there is a very high probability of a strong market move coming.  I don't need Gann, because this gives me time, the square of time, and price all in equilibrium. When it breaks, it goes.【BINGO 收集整理】

Need proof? Well, go back on your favorite currency pair and check it out. In the first quarter of 2005, this filter alone produced 20 trades, 19 which were profitable in USD/CHF. In fact, as I write this, 1 trade is still on from about 3 handles ago. Since I am not responsible for Swissy, I'm not the guy pushing the button, only monitoring it when I'm at the screen [changing stops when needed, etc.]. But, the position is still on.

This filter is so profitable, we increase the size of our trading position when we see it develop and then happen.

When you go back and check it out, you will notice many times how it just misses a move by a few hours. It is an extremely profitable filter.

We also define "same price" as being within 5 pips or so of being equal. Sometimes it turns out the signal is exact, but I don't think you have to split hairs on this. Within 5 pips is good enough for us.

2.)

We do not initiate new currency trading positions based on tunnel trading during the Asian time-frame. Anything between 5pm NY and Midnight NY is ignored for entry of new positions. Positions that are on are monitored as normal, i.e., everything else is the same. We will take profits if fib levels are hit. If we miss a move, then we miss a move. A missed move is just an opportunity cost. Chop-chop in Asia will eventually cost you more money than it is worth.

3.)

News days that can have a significant affect on prices are ignored. That's right, we skip them for entry of new positions. Currently there is only 1 day per month which qualifies, and that is US Non-Farm Payrolls [NFP] which comes at 8:30 am NY time the first Friday of each month. Positions that are on are monitored as normal.

4.)

When the tunnel is very narrow [most of the time], do not just put stop on the other side of tunnel. If you do you get whipsawed to death. Use the hourly charts and the most recent hours of support and res. to make the call.

If you are a newbie to trading, you will find this to be the most troublesome filter. If you are not familiar with trendlines, triangles, flags, pennants, and support and res. levels, then go get the eduation and come back. Simple but necessary advice.

I don't mean to infer that just because you know this technical stuff it's going to be a walk in the park. It's not. Let's make one thing perfectly clear. EVERY model has its vulnerable spot that seem to increase losses. For tunnel trading, this is one of the scenarios. Putting in the right stop is an art, not a science.

5.)

We look for clean moves [1 bar] through the tunnel. This means your into profits almost from the get-go. You will not always get the clean moves. The longer the market stays in the tunnel chopping around, the higher the probability our entry decision will be made on a break of support or res. instead of the tunnel boundaries.

We do not trade minor [contra-major] trend signals in a strong up or down market price trend. If the GBP/USD is in a strong price uptrend, we will not initiate new short positions on a break of the lower tunnel boundary. Why? Because the probability of success in getting past 55 from the ema is not very good. Past history tells us that, so I'm not looking to be the hero here and say "This time it's different." When market comes back through the tunnel on the upside, we will get back in on the long side.

If I have to tell you when the market is in a strong price move, I don't think you have been paying attention to the price movements of late.

In a range-bound market, which we define as a market between 3 - 5 handles [or lower] in a 5 week time-frame, we trade both sides.

Now, that's all we use. Can you use more? Can you invent your own? Can You change some of the definitions? Yes, absolutely. Invent your own filters, use an Elliot Wave filter, anything you think will help your trading.

SUGGESTED MODEL SYSTEM

Do I really need to mention money management?

I didn't think so.【BINGO 收集整理】

At a minimum you should be able to do 3 units to implement tunnel trading. Use the 55, 89, and 144 levels to take 1/3 off at each level. If you can do 4 units, use 55, 89, 144, and 233. 5 units is the preferable level, and you use 55, 89, 144, 233, and let one unit ride until crosses over tunnel boundary or it reaches 377.

Of course, you can make your units any size you want. For smaller traders, a unit size may be 10,000. If you do not have the money to trade 30,000 of something, then I would advise you to save up and come back when you do. If your account has $2,000 in it, you can easily implement tunnel trading with 10k units.

One of the greatest advantages of this model is its flexibility in its design to allow you to choose the level of risk/reward you desire in trading. You can make this as aggressive or as conservative as fits your style. I will give an example of each. These are just examples, I'm not saying you have to do this. I'm only giving you these two to stimulate your brain. In the following day and weeks I am confident you will find an appropriate level for yourself.

Example 1 - Very Aggressive

Tunnel is pivot level for buy/sell. Above tunnel, buy breaks, sell at fib numbers. At 233 an 377, fade the move for retracement. Below tunnel, sell rallies, buy at the fib numbers. Use previous fib numbers in the move as stop loss points. This is very aggresive, and woul be appropriate for very short-term traders who have a time-frame of day-trading.

Example 2 - Very Conservative

Uses basic tunnel system with 12 ema. Only initiates on this signal. Looking for best possible probability trade. Willing to give up more profitability in return for less risk. Trades three units. Uses fib numbers 55, and 89 for 1/3 each. Leaves the other unit on until 233 or market price crosses over tunnel boundary. Allows trader to catch short-term [1-5 day] profit points, and also allows him/her to ride the major trend if one develops.

Like I said, these are just two of an infinite number of risk/reward senarios you can develop using this model. This is not some rigid system, where you have to do this or that. It is adaptable, with no right or wrong answers. This is why many locals from soybeans to bonds to gold and silver, oil, etc. use it. I've seen some people who have transformed this into a model you wouldn't recognize without knowing what tunnel trading offers.

When you get right down to it, once you have adapted it into your own trading style and personal risk model, tunnel trading will give you all you want. Momentum to catch the bigger moves over time, early profit points that allow you to catch short-term movements, and the lowest risk you can possibly have in a trade, because you are only risking 10 -25 pips on each trade. If your odds of success on each trade were 50-50 [they aren't this low], over time you would make a fortune. If you don't believe me, then do the math.

Precisely because of this flexibility tunnel trading is the best model I have ever seen.

                        BY

                       VEGAS

當你編好一套自以為適用的系統後，不要以為這樣就可以等著發財了。實際可沒那麼簡單。除非你“深知”這套系統應該有用，不然則毫無意義。可以利用多種方法，來進行測試你所準備採用的交易策略是否有用。你可以直接應用到市場上實戰來進行測試；可是，萬一測試失敗，結果恐怕會很受傷。你也可以先進行紙上模擬測試。最有效的方法，就是利用歷史資料進行測試。換句話說，就是針對你準備進行交易的市場，採用某段時間的價格資料，測試交易系統的操作績效。幾年前，這類測試的程序非常煩瑣，因為所有的工作都必須依賴人工。現在的情況則不同，我們擁有高速的電腦，某些軟體也具備曆史測試功能。

為何要進行歷史測試？【BINGO收集整理】

但只要經過適當的測試，你就可以了解，系統可能發生的最大連續虧損的程度。也許連續兩個月發生10筆虧損交易，最大損失超過1 萬美元。如果過去發生此種情況，就沒有理由相信將來不會重演。了解一些系統的最糟的情況，可避免實際運用時才發現，自己的資本無法承受或交易風格不願意承受這類損失，而導致被迫中途放棄一套原本可以成功的系統。交易系統確實可能發生連續損失，這屬於正常現象。因此，一定要了解連續虧損的嚴重程度。

歷史測試經常發生的錯誤
在進一步討論歷史測試方法前，我想先花點時間來說明在測試過程應避免的一些常見錯誤。有時，避免錯誤是學習正確方法的捷徑。如果能夠避免錯誤，所作所為自然就會正確。舉例來說，我的蘇非貓並不知道只能在特定的柱子上磨爪子，但經過學習後，明白了它不能在沙發、窗帘、地毯、家具或我的腳上磨爪子。於是，它能夠選擇的地方被侷限到一個地方。現在，我只要能夠想辦法讓它了解，早上5 時並不是舔我的臉、說早安的適當時間，那麼一切就都沒有問題了。稍後會闡述正確的做法，不過現在先來講講一些坏習慣。

不知道如何評估測試結果        歷史測試完成後，最常見的錯誤就是不知該如何評估測試的結果。除非能夠作出有效的評估，否則，你根本無法判斷系統的功能。一些人特別重視測試記錄中的最高凈報酬或最高勝率，可一旦系統的最大連續虧損過高，上述資料就沒什麼意義了。評估過程必須綜合考慮一些東西，如：交易筆數、每筆交易的平均獲利、連續發生虧損的交易筆數、最大單筆虧損、最大單筆獲利、平均交易筆數、報酬分配狀況等等。只有綜合考慮以上因素，才能判斷出一套交易系統是否適用，或判斷該系統相對於其他系統的操作績效。

曲線套入與過度最佳化        測試結果是否經過曲線套入，是評估系統績效所需要考慮的重要因素。什麼是曲線套入呢？簡單地說，就是根據資料狀況來設定系統策略。這是交易系統建構與測試過程中都可能發生的問題。如果你發現價格資料呈現明顯的趨勢，然後才建構一套買進和持有的系統，這就是曲線套入。建構系統的過程中，交易者可能特別注意走勢圖上適用其預定策略的資料，但有意無意之間卻忽略其他不適用的資料。這樣建構的系統，很可能只適用於某類市場狀況；如果沒有採用其他曆史資料進行測試，這套系統顯然不該被採用。過度最佳化則是另一個因素。系統建構過程中，如果對於參數設定要求十全十美，就可能發生過度最佳化的問題。舉例來說，一套由兩條移動平均構成的穿越系統，建構者可能會測試所有可能的均線組合，然後挑選一組績效最佳的。這類系統未必適用，就如同精心培育的溫室中的花朵，往往都禁不起殘酷環境的考驗。所以，測試過程中，不要為了取得較好的測試績效而不斷調整系統參數，因為這類系統通常都不適用於未來。【BINGO收集整理】

不質疑系統        不懷疑系統的測試結果，也是一種錯誤心態。某些交易者看到不錯的測試記錄，就覺得很滿意了。實際上，使用者應嘗試尋找系統的毛病或不正常的地方，觀察是否有曲線套入的傾向，因為在事前了解這方面的問題，總要比實際遇到的好。系統的測試績效很好，可能只是由一兩筆交易造成。如果遇到不同的市場狀況，這套系統還能發揮測試過程的績效嗎？測試過程對於滑移價差的假設是否合理？測試過程的交易樣本數量是否足夠？如果你不嘗試質疑系統的績效與缺失，實際運用中就可能遇到不必要的麻煩。對於一套看起來似乎沒用的系統，不妨也抱著這種心態，研究其問題所在，看看自己是否能夠進行修改。檢討一些無效的策略，也可以提高交易知識。任何假設或理論，都需要進一步研究與質疑，才能知道它是否成立。