Sari la conținut
  • intrări
    18
  • comentarii
    51
  • citiri
    3.433.354

To Trade or Not To Trade the Time (part 1)


tradelover

2.333 citiri

Va place aici?  

9 membri au votat

  1. 1.

    • Da
      9
    • Nu
      0

Well, since I am participating in some "Automatic Trading Contest" I've learned a lot of things about trading the time. Before I was sticking to my brokers, and never went to others, but the contest opportunity forced me to trade other broker's quotations, showing me at least few strange things. This blog entry is not targeting the scalper, momentum trader, nor the investor, but the medium range trader, like a swing trader, or one that intensively use indicators like Stochastic or RSI, on H4 or D1 time frames.

 

What is the point? Can be the price "broker dependent"? Well, not. Except few small fluctuations, spreads, some spikes of "criminal" brokers, generally the price will follow same quotation lines, independent of the broker. If the price curve is different for one broker, then few arbitrators will jump in with lots of money, targeting fractions of pips profit of it, gaining money, equilibrating the path of the price, and bringing loses for the broker. So, the interest of the broker is to be fair, except few occasionally stop hunting or FPI discrepancies. If you have such a broker that is hunting your stops, better change it. They are less and less now, because the competition is hard, there are more and more "good brokers". Here I will consider that all brokers are good. But they, and we, have all a big enemy. That's the TIME.

 

I said many times that scalping (I define scalping like placing lots of orders with high lots qty, short SL and shorter TP, other kind of trading is not scalping) is not the way. Some scalpers make money, but they are just the exception. The most traders that are profitable would be medium to long term traders. This means they analyze the fundamental factors first, then they use indicators on H4, D1 (or even higher TF, as W1 or MN1, for long term investment) to establish their entries and exits. As we are mostly swing traders, our problem now is H4 and D1. Can the price fluctuate between the brokers, over this time intervals? No, as I said, they can not. At least, not significantly. But how many of us are using solely the price in their trading? I am sure: NONE. All of us use indicators. All of us use MACD, moving averages, Stoch, RSI, CCI. And ALL these indicators will differ from a broker to another, when we place them on H4 or D1 charts. Why? As I said, the TIME. The great king of time...

 

Different brokers are on different time zones. Few brokers close a D1 candle at 0:00 GMT, other will close the day when NY session closes. The most of the brokers will close the day according with their local time. And that makes all the difference. Sure, everyone of you knew that brokers have different time zones. But have you ever asked yourself how much one indicator can change if one broker is closing his D1 candle one or two hours before the other?

 

When I start testing MetaQuotes contest (demo) account I found out that all my experts produce different results compared with the results produced on my original brokers. There was a time difference of two hours between my broker's server and MQ server, and all H4 indicators (most of them proprietary) looked completely different. I had to adjust them and to change many settings to be able to join the competition. Then surprise! After few weeks the time difference between my broker and MQ changed again. It was now just one our difference. I don't want to say that could be the reason why I am losing in the competition :D, far away from me. I am losing because the expert was stupid, or unlucky, or bad programmed, and other people are much better. That is a different story. What I want to show is the conjuncture that lead me into playing with the time. MQ (that is not a broker, but a Russian company selling brokerage software, and they provide a testing server where you can trade demo, but you can never trade for real money) had Daily Saving time, that payed out at the end of October. They moved their server time to GMT+1, from GMT+2 before. My broker stays on GMT+0. I had the opportunity to see in a very short period of time, the behavior of H4 chart indicators, with 3 (from 4 totally possibilities) different time zones. And that was really painful. I spent many days and nights before the competition started, just to understand why my indicators show different results, to correct the problems and to optimize. It was very easy to spot that: when H4 candle of my broker closed, the H4 candle on MQ was just half developed, because of 2 hours difference, and viceversa, when MQ was closing their H4 candle, my broker's H4 candle was just in the middle of the race. Easy to spot, but how to make it working? Anyhow, I found somehow a "patch" for this story and I went on. That is not the point. But later, when the DST took place, I had everything very fresh in my mind. Now as the time changes, I also saw different values of the indicators, and moreover, during my real account was losing a small sum of money, the competition account was gaining huge, and viceversa, the huge loss of the competition account was balanced by a small gain on the real account. They were traded by the same expert advisor, only with a different risk factor (for the competition - demo money - I raised the risk to the sky, otherwise no chance to win, but the real account has a normal risk factor - below 1% - and won't take any bets against the swap). So, how in the seven seas can that be? They were suppose to gain both in the same time, or to lose both in the same time. Different amounts, but same direction. Or, what I was witnessing was a strange situation, in which one account was going down during the other was going up, and vice versa.

 

Well, to spot the answer is easy, again: The time. I was using stochastic in the expert, and one of the most sensitive indicator to such time-shifting is stochastic. Why? because the Stoch indicator is percentaly computing the closeing price in relation to high-low difference. Supposing your candle has a high price of 1.20 and a low price of 1.10, if the close price of that candle is 1.123, then the Stoch of that candle is (1.123-1.10)/(1.20-1.10)*100, that makes 23. What you see on your chart as "Stochastic Indicator" is just a moving average of all these "candle-stoch" values. Now let's suppose that after this candle the price is pulling back, with exactly same min and max, and it close somewhere between the two values. We have a new candle in reverse direction, with the same min and max, and if we compute the stoch, we get some (not important) value between 0 and 100. The idea behind Stochastic is that in an uptrend the candles tend to close near the max point (an average stoch value over 50), and in a downtrend they tend to close in the lower half (an average stoch value under 50), otherwise it would be no trend in that direction.

 

Now let-s imagine we shift the time with half of the TF. For example if we work on H4, we shift the server time with 2 hours. The new chart could have started the former bar somewhere lower then 1.20 (say 1.15, the second half of the first bar in the first example), it would go also down to 1.10 (that is minimum for this period) but for the next two hours it will raise (the first half of the second bar in the first example), say to 1.137, and close there. Computing the stochastic for THIS bar, we get (1.137-1.1)/(1.15-1.1)*100=74. Woowwww! There is exactly the same currency pair, the same quotation, the same time frame, the same index of the bar, the same PHYSICAL (real) time. But the server time shifted two hours. And one calculus placed us close to oversold, and the other placed us into the overbought.... Of course, in reality such differences are leveled by the averaging of the stoch indicator with %D period, and they will never be so big as in our (forced) example. But there are plenty of differences. And my concern now is to see if we can use this fact in our advantage.

 

Can our system rely on such indicators? We will try to analyze the effect of time shifting, with few images. We are on H4 chart, and we paint first the "standard" stoch indicator. Each H4 bar is made by 4 H1 bars. We divide the H4 bars (by going to H1 chart) and regroup them in all 4 different ways. That is if you have a broker with GMT server time (or GMT+4, or GMT-4, or GMT+8, etc) then first H4 candle will start at GMT, next one will start at GMT+4, etc. If you have a broker with a server time of GMT+1 (or +5, or -3, etc) then first H4 candle starts at GMT+1, next one at GMT+5, etc. Obviously, we have 4 different choices for all 6 H4 candles that make a day. That is, we have 4 different categories (classes) of H4 traders! If John has his broker in one class, he will see a stochastic indicator on his H4 chart, but this indicator will look different from Mary's one, who has a broker in a different class.

 

How different? Well, not too much. Even if corespondent candles give candle-stoch values quite different, they will be leveled by the moving average of %D. But the remaining, is still quite significant. Let's have a look to the "artwork" below :D

 

blogentry-1272-1196148881_thumb.jpg

 

Fist indicator is "classical" stoch, built in MT4, that is computed, of course, when each H4 candle closes, according with your broker time, at his server time plus 0, his server time plus 4, and so on. Next 4 indicators are computed from H1 chart, by grouping the bars 4 by 4, in all 4 possible modes. One of them should be exactly the same as the "normal" stochastic from the higher window. In our case is the first one, denoted "Stoch+0". So the first two have to be exactly the same, if we did our math right. In fact, they are the same for almost all bars, but sometime there will be a small difference on the first candle of the week (weeks are marked with period lines) and on the last candle of the week. The difference come from the fact that H4 bars at the beginning/end of the week are usually not "complete bars" (they sometime contains only 2 or 3 hours of price action, due to market closing time on Fridays and market opening time on Mondays, or Sunday night, depends on your broker's meridian). We will ignore that small discrepancies now. But moving down tot the other 3 windows, denoted "Stoch+1", "+2" and "+3", we can spot many other differences. Are they important? It does not seems to be a big deal. Of course they are!

 

Sometime we can spot a hook-situation on some of the indicators, but the others have no hook. Sometime one is leaving the OB/OS area, but the others not. We can easily imagine that different traders around the globe, looking to THEIR H4 stochastic will have different opinions about where the market is going, if the market is OB, or OS, if the resistance line is going to hold or not, moreover, some of them will see divergences, some not.

 

Traders using different time zones having different opinion about the market? C'mon! That is since-fiction. Well... it is not. Let's superpose all this indicators on the same window. We will use the standard stock, and the last 3, as we don't need Stock+0 anymore, its purpose was just to check if our math was right. If we suppose that our broker time is "convenient", in such a way that opening of a H4 candle coincides with opening of London or NY session (that is our case in fact) then we have the following scenario:

 

Stock+3 are the brokers class (trader's class) trading the beginning of the sessions, they close a H4 candle one our after big volumes starts to flood the market. They are followed by Stock+2, then Stock+1 classes, that will close their candles in the middle of the session, and respectively at 3 hours after. Then we, the lucky ones, with Stock+0 (or standard stochastic built in MT4, if we got a broker with a "lucky" time zone) came at the end. Have you ever spotted that the big movements are at the beginning and at the end of the candles? Look to the volumes. Look to the lengths and shadows of the candles. So now I am asking you, who of the for groups are the "luckiest" one? We say in Romanian language that the guy who laugh last will laugh better (no idea if any English equivalent).

 

Now let's imagine that we are "big money" on the market. The BIG Sharks. What we would do? We have the knowledges, we have an army of people watching the charts and digging for economical news. Or (why not?) spreading news... Bullish news... So let "the papers" buy. Yeaaaa, let them buy! Then we sell. How the things are really going on? Well, exactly in the same way. Every time the trend reverses, say from bullish to bearish, who will make the most of the money? Of course, the guy who sold last (or first, in the new trend). Who will lose most? The one who bought last. Who is making money in forex? The profesionals. The Sharks. Who is losing? Us! The beginners. Every trend is a fight between beginners and professionals, every candle is a fight. See Alexander Elder, Trading for a Living. First two chapters (about trading psychology) are the best trading material I ever read. The rest, trading strategies, all authors talk the same bullshit, trend lines, indicators, rules, bla bla... even some of them talk about planets and zodiacs in trading... Grrrr...

 

So, every trend has a big shark in his peak, and has a big shark in his deep. They trade last. They laugh last. Going back to our "superposed" indicators. What they will do in such situations? A H4 candle starts. People start joining in. They are buying. One hour passed. Stoch+3 closes now. It contains 3 hours from the former H4 candle and 1 hour from the current candle. Ho it will look? It will be up. UP, UP, UP, UP!... Others are selling to cover. Another hour passes. Prices come back. Stock+2 and Stock+1 come back. Then the big shark put his fins in the pool... Only the fins... Big sharks never jump in the pool, they won't like all the water to go out. So Stoch+0 will close the H4 candle. And it will close it down. DOWN.

 

As long as the trend is strong, there is a consensus. All Stoch+X go together. But as long as the opinions of the traders start shaking, they split. Stoch+X split too. And that is the moment for the big shark. That is the fragile equilibrium that he can use in his advantage, to push the market toward his pocket. Don't believe? Look here:

 

blogentry-1272-1196153034_thumb.jpg

 

We see a very clear split in the beginning. There is agitation. We are in a long and strong up-trend on cable (check the history) and that is the end of the 4th correction already. Will the trend end? Suddenly opinions start to vary. Will the trend continue? Some other correction will follow? On that two H4 bars, different people trade different opinions. But the big shark is on the fence. Sometime they give you a small bite, just to feel the taste, so they can catch you later. And this is clear - Stock+0 (yellow line) is on the middle. The sharks are waiting, or they did a small "sell to cover" from the previous trend. Then pay attention to the zone A. What do we see there? All Stoch+X lines are strongly overbought. They are close to 100, and no tremor. In that moment ALL the traders from ALL time zones saw the OB condition. The impulsive wave is gone. Then all the stocks went down, making a few hooks on the way. Strong sell signal. Moreover, the peak on June 5-6 allowed them to set a tight SL. So, here we are, the corrective went 500 points down. Then Stoch+X lines split again. Traders are again undecided, if the trend will resume, or the correction will continue and the trend will reverse. Some push down. Some push up. Stoch+3 (red line) is pushing down. Stoch+0 (yelow line) is up. Professionals are pushing up. Where do you think the trend will go? He, he, beter buy to cover, if you took the sell on A. And here we are with about 500 points profit. If you did not buy to cover yet, then the divergence (yellow thick line) is a good time to stop and reverse to bullish. Buy to cover here, take the sell profit (about 400 pips) and buy again the divergence, especially because - guess what? - the stocks are splitting again, now in a different direction. The rightmost part, red thick rectangle, Stochs split again. There is a signal that people started to sell to cover, or trade an eventual new bearish trend. Better close you buy, if you have one, or move the stop tight.

 

This situation is not singular. Of course, I put the salt an pepper on it. But we can find such discrepancies on every currency pair, and with every stochastic period. I am not quite sure yet what can we get, putting all together. I certainly need more time for experimenting, and a little bit of forward trading for more "credible" testing. Up to then, let's see few more images.

 

blogentry-1272-1196157677_thumb.jpg

 

Kind of same situation, just a bit forward in the future, July 4th is a good opportunity to drive the market crazy, some holiday, low liquidity, sell first and all the freaks will jump to sell. Then buy, and pocket a lot of money one month later. But for the clever trader who doesn't went short, the OS condition, combined with Stoch+X discrepancy is a good chance to buy when the Stoch+X cluster is leaving the OS zone. Again, a divergence and drill through the support line tells him when to cover, or when to reverse. Some cautious or medium-range trader could go out on the divergence, making at least 150 pips more.

 

Switching to Euro, the picture now is different. Dollar sentiment is bearish, the price can't leave the channel, and can't break the historical support line. Adding Stoch+X to the equation, they "are getting nervous" every time when the price is hammering the support line, and every time there is a hook and a rebuild (leaving OS zone) of the Stoch, telling us that we have to add to out (eventually) longs. Where that will stop? God knows...

 

blogentry-1272-1196158844_thumb.jpg

 

Note that we could make some money also trading the small correction at the beginning of the chart, from exactly the same (Stoch+X) reason, but told now upside-down :D

 

The last chart is a bit more complex. We are after a long trend, and there is a big indecision, signaled also by the price (almost all traders who went in profit after the long trend, sold here to cover). Stoch+X are dancing crazy. There would to risky to go short here. Except of the small triangle, there is no short signal. And the triangle is broken too close to "no zone" area. So, we have to be cautious, and eventually, if we have any open long position, sell to cover and guard the profit. Then there is a pull down, and an almost perfect hook on all the stoch indicators.

 

blogentry-1272-1196159409_thumb.jpg

 

The former deep from the end of May is not beaten, and many people would take that as a long signal. But going long is - again - risky here: there is no place to set the stops. Moreover, the indicator is taking off, together with the price, is close to OB and it starts going crazy again. Well, there is no way to buy. If there is to be something, then it can be only a SELL. The former top provide a nice resistance level to place our stop behind (above) and the last drop provides a perfect "fixed point" for a Fibo extension that help us to establish our target. I would really SELL somewhere inside of the yellow square, with a stop at the blue dashed line and a target around 1.3400. If I am unlucky, I have a tight stop. But if I am lucky and the price drops as it dropped, there are plenty of signals telling me to go out there, before the rebuilding of the trend starts. Among all of these signals, Stoch+X is getting crazy again....

2 Comentarii


Comentarii Recomandate

  • Moderators

This is a part of more complex tool, including the daily shifts of the stochastic, beside of H4 shifts exemplified above. I did the same thing for daily charts, but painting 4 stochastic indicators, one is the default one, as given by my broker, and other 3 correspond to the Asian, London and NY sessions, they are also daily, but shifted with the right amount of hours. It make no sense to do this for hourly, as all brokers start the hour at XX:00 sharp, and also is quite irrelevant for weekly charts. But for H4 and daily, THERE ARE discrepancies that CAN be traded!! Yet, I don't know how... Putting all together, this tool is very complex and far away to be complete, and I am still working on it, to complete it. I did not make any real money with it, but it shows good results on demo. I am trying to do something like "stochastic arbitrage" :D, starting from the idea that if the stoch lines spread, they have to come together again somehow, in the future. But the results are.... not significant, as it was discussed before in The Economist's blog, there is no potential to make any arbitrage with only one instrument, the market is laughing at you when you are less expecting it... You need a second instrument to hedge, otherwise it appears like no profit can be made.

 

Regardless of good or bad results, this tool will not be distributable. If it is good, I will keep it for myself, if it is bad, make no sense to distribute it. People here know: if I want to give something for free, it is posted already. I don't sell tools for money yet, so no meaning to ask :D

 

But I gave you the idea, you can do your homework. If you did not started already to program MQL, maybe this is a good opportunity :P I have the advantage of being a good programmer, but from trading point of view, with this method I am exactly in the same point as you are: experimenting. This is not a "sandpapered" method, just an "axed" one yet...

Link spre comentariu
Vizitator
Adaugă un comentariu...

×   Alipit ca text avansat.   Alipește ca text simplu

  Doar 75 emoji sunt permise.

×   Linkul tău a fost încorporat automat.   Afișează ca link în schimb

×   Conținutul tău precedent a fost resetat.   Curăță editor

×   Nu poți lipi imagini direct. Încarcă sau inserează imagini din URL.

×
×
  • Creează nouă...

Informații Importante

Am plasat cookie-uri pe dispozitivul tău pentru a îmbunătății navigarea pe acest site. Poți modifica setările cookie, altfel considerăm că ești de acord să continui.