Multi-Lot Multi-Timeframe Position trading | Page 51 | Forex Factory

Options

Go to
Permalinks
Print Thread

Similar Threads

building multi-pair multi-timeframe scanner 48 replies

Multi timeframe multi chart 13 replies

Timeframe inside current timeframe - is that possible? 2 replies

Seeing higher timeframe indicators on lower timeframe 4 replies

BUY run profit 1 lot, SELL LIMIT 2 LOT down again n BUY 1 lot - EA Reqd 5 replies

Attachments: Multi-Lot Multi-Timeframe Position trading
Exit Attachments

Multi-Lot Multi-Timeframe Position trading

  • #1,001
  • Mar 2, 2015 12:23pm Mar 2, 2015 12:23pm
If we were to take a short trade in EU and set our stop just above the Lower high I've labeled as Price Peak #1, we have a fairly good chance that price will violate that stop before the trade can run to glorious profits. That is to say, price would temporarily travel above that previous Price peak before continuing on it's way along the down trend. The probability that price will violate that previous lower high in price varies depending on the time frame we are looking at volatility of the pair and to a lessor extent other factors, but we won't be able to trail a stop this close many times before price violates the stop. after 3, or 5, or 10 stop trails, price will violate the stop and kick us out of our good trade and we'll give back some profit in the process. We need a better method, but one simple enough we can quickly set and adjust our stops. I like being able to do all my trading calculations in my head, and my head isn't as snappy as it once was, so I want to keep this simple.

Certainly volatility is a big factor here. The more volatile the pair, the more likely we would get kicked out of a good trade early. Many traders use the ATR average true range to account for volatility, but it has some drawbacks. ATR usually averages volatility over some previous number of bars, like 10 or 20 or so. But price volatility cares little for how we count bars. In the example above, volatility was fairly low for the previous 40 bars, then extreme for the 10 bars before that. If you take the average volatility of the previous 10, or 20 or even 40 bars, you'd conclude that a very tight stop would be fine. Even if you average in the volatility of the sharp drop 50 bars ago, the result is still low due to the recent majority of low volatility bars averaging in. We need to account for volatility, but the ATR isn't a good enough measure of recent volatility for our purpose. Discussion, questions, comments?
  • #1,002
  • Mar 2, 2015 12:27pm Mar 2, 2015 12:27pm
Quoting Graviton
I'll move on and if anyone has any comments they can jump in at any time. First of all, allow me to clarify our definitions a bit. We have already adopted the discrete definition of a down trend as a series of lower highs and an up trend as a series of higher lows. Let's also adopt a discrete definition of a higher high and a lower low. There can be lots of little zigs and zags in price, so we want to adopt a definition that will always produce the same results. For this presentation I'll call a price peak one that occurs at about the same time...
I like your Stoch G. Had not seen this one before. Much better than 8,3,3 on H1.

AG
Do not dwell in the past, do not dream of the future, concentrate the mind.
  • #1,003
  • Mar 2, 2015 12:29pm Mar 2, 2015 12:29pm
Hi G. and @all,
this current discussion you are opening is exactly what I am thinking about to get a suitable solution.
My opinion: If 1 out of 10 cases ( as you said) the previous swing high/low where the stopp is set is hurt and in 9 times the trend continues without catching the stopp its ok for me.
There is nearly almost the chance to reentry the trend on daily/weekly timeframe.
I was looking on hundreds of charts to find a better solution but......
One alternative perhaps: switching SL only if there is another second new swing low/high- this method survives the "head fake" but at the end of the trend it gives back a lot of pips to the market.....
My dream is to "mechanize" a three-position-entry:
position 1 SL set to BE after the first swing high/low has happened and not moving SL for a minimum of 1000 pips catching monthly trends.....
position 2 SL set to BE after the first swing high/low has happened and moving SL to any new swing high/low (tight trai stopp) or moving SL after the second new swing high/low has happened (wide trail stopp) for catching most of the swings...
position 3 SL set to BE after the first swing high/low has happened and setting a target stopp at the next PPZ or RN or support/resistance for catching short-term objectives....
Keep discussing
Have a nice day
  • #1,004
  • Edited 7:16pm Mar 2, 2015 6:15pm | Edited 7:16pm
One simple way to incorporate high volatility in our stop setting without having it diluted by recent low volatility bars is to look at the size of the recent adverse price excursions and base our stop setting on the recent worst case. For instance, in the EU example previously shown we can simply look at each of the three previous Lower High Price Peaks and place our stop outside the most extreme of the three. This has several additional advantages such as automatically adjusting for recent worst case volatility and setting the stop outside any recently tested s/r. In most cases this will produce a reasonable stop somewhere near the ATR or ADR of the pair. In the case of the EU example though, due to the recent large fall in price, the stop would be very large, almost 200 pips between current price and the price labeled as Lower High Price Peak #3. Since that price move was in the direction of the short trade we would be taking and not against it, we have to question whether such a large stop is really required when trading in the direction of the major trends. Some common sense is always required in trading, but fortunately for me, not much.

We should also note that we aren't looking for a definitive solution to the worst possible case here. We are looking for a statistical solution that will make good profits for reasonable risk in most cases. As stated earlier, there is no practical limit to how large a price excursion could possibly be. So in other words, we are looking for a stop that will take us out of a trade protecting most of our gains in the rare event of a major retracement or an actual reversal in trend, while keeping us in the trade to enjoy future gains through the smaller retracements that occur much more frequently.

Looking through the first set of charts, or any set of charts you choose where price is trending in all major timeframes, we can see that price excursions that retraced back above most recent higher low price peak were rare, perhaps only around 1 in 5 to 1 in 10. Price excursions that retraced beyond the larger of the previous two higher low price peaks were even more rare, perhaps 1 in 15 to 1 in 30 or so. I don't have good data for the occurrence of a price excursion above the largest of the previous 3 higher lows without a deep retracement or reversal in trend occurring (a good homework assignment for the reader), but I'd think that would happen in less than 1 in 50 occurrences. If we are trailing stops every time a new lot is put on and have spacing of 50 pips between lots, that would allow us to ride out a trending trade of about 50 stop trails of 50 pips each or 2500 pips without being stopped out prematurely. That sounds optimistic, but even if we can stay in a trend half that long, and not give back more than the worst case 200 pips shown in the previous EU trade when the trend comes to an end, we will profit by over 1,000 pips in a long trending trade. I think we have the beginnings of a solution here Comments?
  • #1,005
  • Mar 2, 2015 7:45pm Mar 2, 2015 7:45pm
Quoting Graviton
I think we have the beginnings of a solution here Comments?
What TF are you concentrating on for the trail? The H1 as you show on the E/U example?
  • #1,006
  • Mar 2, 2015 9:15pm Mar 2, 2015 9:15pm
Quoting spyderman
{quote} What TF are you concentrating on for the trail? The H1 as you show on the E/U example?
Yes, I'm just looking at the H1 at the moment, but due to the fractal nature of price movements, the same effect should apply to other timeframes within reasonable limits.
  • #1,007
  • Mar 2, 2015 10:56pm Mar 2, 2015 10:56pm
I realize some of my terminology may be confusing with made-up terms like lower-high price peak and such. If anyone has a question let me know and I'll try to explain differently.
  • #1,008
  • Mar 2, 2015 11:44pm Mar 2, 2015 11:44pm
Here is how I put my SL and TP usually ... using statistics! :-) The example below is using weekly stats.

Below is an example of analysis I did for a pair if its an up month, week, or day. Lets say trend is up, and we don't know if current week will be up or not (lets assume it is). If price opens above the weekly pivot, I see there is a 0.8 probability of L < pivot for the week, but only a 0.13 probability of closing below the pivot. I can then put a limit order to buy at the Pivot.

Now for SL, its only a 0.18 probability of closing < S1 for an up week, so that's a good area for SL. For TP, usually I trail, but for stats ... P(H>R1) = 0.7, but P(Close > R1) = 0.45. So we know statistically price is over the R1, but won't close over R1 (not probable). This is a good place for TP.

I hope you get the idea. ;-)

EZ
Attached Image (click to enlarge)
Click to Enlarge Name: Screenshot2.png Size: 29 KB
  • #1,009
  • Mar 3, 2015 3:18am Mar 3, 2015 3:18am
Thanks Graviton, that is an interesting way of looking at stop placement.

Do you think there are differences between pairs? For example the more liquid pairs such as EURUSD seem to move much more slowly and are more prone to pull-backs.
  • #1,010
  • Mar 3, 2015 3:26am Mar 3, 2015 3:26am
Quoting Graviton
... but the ATR isn't a good enough measure of recent volatility for our purpose. Discussion, questions, comments?
Hi Graviton,

sometimes I am a bit confused when I have the feeling you change your opinion about something suddenly. I think AlbineGrande was also surprised that you jump into stochs as you usually talk about RSI and Moving Averages. I also remember you were discussing earlier the advantage of ATRs, now you say they are not as good. Actually do you use them or not? I just would like to be clear on it.

I really liked your awesome post on having SL higher than the 3rd last lower high going back from the recent price. I did not read something about a method like this and I believe it has to work great in reality.

Thanks for your input,

IBO
  • #1,011
  • Mar 3, 2015 5:29am Mar 3, 2015 5:29am
Quoting ezcurrency
Here is how I put my SL and TP usually ... using statistics! :-) The example below is using weekly stats. Below is an example of analysis I did for a pair if its an up month, week, or day. Lets say trend is up, and we don't know if current week will be up or not (lets assume it is). If price opens above the weekly pivot, I see there is a 0.8 probability of L < pivot for the week, but only a 0.13 probability of closing below the pivot. I can then put a limit order to buy at the Pivot. Now for SL, its only a 0.18 probability of closing < S1 for an...
Thank You. interesting idea.
  • #1,012
  • Mar 3, 2015 6:24am Mar 3, 2015 6:24am
Quoting Graviton
{quote} Yes, I'm just looking at the H1 at the moment, but due to the fractal nature of price movements, the same effect should apply to other timeframes within reasonable limits.
But for trend stoploss placement, wouldn't we have to pick a specific time frame to go off of. If we were selecting the highest of the last 3 peaks as the stop and those peaks were defined as a crossing on the stochs, then that crossing would obviously vary depending on the TF.

A stochastic cross as you've described on the 1H tf might not even be a blip on the weekly.
  • #1,013
  • Mar 3, 2015 8:18am Mar 3, 2015 8:18am
Quoting ezcurrency
Here is how I put my SL and TP usually ... using statistics! :-) The example below is using weekly stats. Below is an example of analysis I did for a pair if its an up month, week, or day...
Yes, a valid analysis. Also, lots of people watch pivots, which may sometimes make them self fulfilling. I often consider pivot placements when deciding where to place stops. For instance, in my discussion of a stop placement for EU, there was a possibility of needing a 200 pip stop to enter, which is too much risk for me. A pivot and R1,R2 placement might convince me that such a large entry stop isn't necessary.
  • #1,014
  • Mar 3, 2015 8:32am Mar 3, 2015 8:32am
Quoting lemna
Thanks Graviton, that is an interesting way of looking at stop placement. Do you think there are differences between pairs? For example the more liquid pairs such as EURUSD seem to move much more slowly and are more prone to pull-backs.
Yes, I'm sure different pairs have different characters. The biggest difference is the different volatilities of the different pairs. That must be considered when placing stops and managing trades. The method I'm exploring takes volatilities into account, as do most good methods. Some pairs tend to trend much better than others, but that is taken into account with a multi-time frame trend analysis. The result of that trend analysis means I wind up trading the same few pairs most of the time since they are the only ones that trend reliably.
  • #1,015
  • Mar 3, 2015 9:32am Mar 3, 2015 9:32am
Quoting IBreakOut
{quote} Hi Graviton, sometimes I am a bit confused when I have the feeling you change your opinion about something suddenly. I think AlbineGrande was also surprised that you jump into stochs as you usually talk about RSI and Moving Averages. I also remember you were discussing earlier the advantage of ATRs, now you say they are not as good. Actually do you use them or not? I just would like to be clear on it. I really liked your awesome post on having SL higher than the 3rd last lower high going back from the recent price. I did not read something...
Well, I suppose I always try to keep an open mind about everything in trading. The market constantly evolves and my trading is dynamic, evolving slowly with the market. It is an evolution rather than a revolution though. Things that worked well 10 years ago may not work as well today. I'm only using stochs for a very specific purpose here. They give me a consistent method to pick out peaks I want to count. There are lots of little zigs and zags in price, but I only want to count major peaks for this count back stop setting method. I'm sure many methods would produce consistent results, but that consistency in picking peaks is my only concern here. I could probably pick these peaks out by eye just about as well, but using the stochs confirms my eyes are seeing what I think they are.

I do always consider the ATR of a pair when placing my trade. It gives me a good idea of what a reasonable range for a stop should be. I know off the top of my head for example that the daily ATR of GJ is twice or more that of EU. My stops typically fall within a range of 0.5 to 2.0 X the daily ATR. If my count back method showed a stop greater than 2.0 daily ATR, I'd have to consider if that's too much risk. Generally I'd think it is. If I show a stop of less than 0.5, I'd have to question if that is really enough to survive minor retracements. But, my point here is ATR can give less than optimum results for two reasons. First, it does not consider recent s/r's, so if the daily ATR is 80 pips, but there was a major resistance just tested at 90 pips, I think a stop on the other side of that recently tested resistance is worth the extra 10 pips risk.

The second reason to look for an improvement over just using ATR is that ATR is just an average over a fixed number of bars, but what we are really interested in is the possibility of an excursion of price that would extend beyond a fixed stop, and then quickly return to the direction of the longer term trend. I think I can demonstrate statistically with a high level of confidence that event is less likely if I count back two or three previous price excursions and place the stop on the other side of them. I don't really have a good way to estimate how much I need to add to ATR to achieve that same level of confidence. So, for example, am I half as likely to hit a stop at 1.5 X ATR than at 1.0 ATR? I might could do some complicated statistical analysis and answer that question, but I can very easily confirm that by counting back 2 previous price peaks that I am less than half as likely to violate the stop as just placing it on the other side of only the most recent price peak.

I don't want to confuse anyone here. I am still performing my multi-timeframe trend analysis before entering a trade and I am still scaling in and taking profits as previously discussed. This is really just an attempt to improve the results of those initial entries and allow a reasonable trailing of stops during the course of the trade. Usually, trailing stops too tight will kick me out of good trades way too early, but perhaps there is a better way to do it that will not increase risk over the broad average of trades. Thanks for the good questions. Please let me know if further clarification is needed or if you'd like to discuss these issues more.
  • #1,016
  • Edited 11:44am Mar 3, 2015 10:35am | Edited 11:44am
Quoting spyderman
{quote} But for trend stoploss placement, wouldn't we have to pick a specific time frame to go off of. If we were selecting the highest of the last 3 peaks as the stop and those peaks were defined as a crossing on the stochs, then that crossing would obviously vary depending on the TF. A stochastic cross as you've described on the 1H tf might not even be a blip on the weekly.
Yes, you have quickly seen an open question in using this method. The selection of the H1 for my example seems rather arbitrary. Why not use the 30M to further reduce risk? Why not use the H4 to increase the likelihood that we would stay in a good trending trade longer? In the case of the EU example, the H1 yielded a risk that I was already willing to take, about 60 pips for the larger of the previous two price peaks. I wasn't willing to take the third price peak back as that would have required about 200 pips stop, so I found a good reason to rationalize that one away. Essentially that I am trading in the direction of the longer term trend and that move was in my direction and not against it.

Using the same method on the H4 would also have yielded a stop of about 200 to 250 pips and that's just more than I wanted to risk, so my options were to either skip this trade as too risky, or use the 2 peak count back which should keep me in the trade long enough for the longer term trend to take affect in more than 9 out of 10 times trying such an entry. But if the H4 or 30M charts offered the entry I wanted to take anyway, say a 50 to 100 pip stop, then I'd have taken the entry there instead.

In fact, I did take that EU short trade. Stop is still resting above the second count back peak I identified on the EU H1 chart, and the trade is positive at the moment. There has now been a new lower-high peak established, so my stop is now above the third count back peak that I wanted all along. Note that this entry is the first of my 12 step sequence, so even if it is profitable I'll take it off at +50 pips and then use that profit to try to get two lots on at low risk. So, I'll have to go through this again to ultimately get a good long running trade on, if that's even possible. Thanks for the observation.

G
  • #1,017
  • Mar 4, 2015 2:23am Mar 4, 2015 2:23am
Quoting Graviton
{quote} Well, I suppose I always try to keep an open mind about everything in trading. The market constantly evolves and my trading is dynamic, evolving slowly with the market. It is an evolution rather than a revolution though. Things that worked well 10 years ago may not work as well today. I'm only using stochs for a very specific purpose here. They give me a consistent method to pick out peaks I want to count. There are lots of little zigs and zags in price, but I only want to count major peaks for this count back stop setting method. I'm sure...
Thanks G for defining the issue in details. No further questions there.
  • #1,018
  • Mar 4, 2015 2:26am Mar 4, 2015 2:26am
Quoting Graviton
{quote} Yes, you have quickly seen an open question in using this method. The selection of the H1 for my example seems rather arbitrary. Why not use the 30M to further reduce risk? Why not use the H4 to increase the likelihood that we would stay in a good trending trade longer? In the case of the EU example, the H1 yielded a risk that I was already willing to take, about 60 pips for the larger of the previous two price peaks. I wasn't willing to take the third price peak back as that would have required about 200 pips stop, so I found a good reason...
I think when we decide which TF we use to look look at the last 3 peaks it is also important to look at the average trade length. If the trade goes usually between 1Hour - 10 days then the 1H TF looks good to me. If we hold our trades a lot longer and have wider SL in place then I would consider to check those peaks maybe at the 4H or Daily TF. Everyone can adjust this to his personal preference.
  • #1,019
  • Mar 4, 2015 3:27am Mar 4, 2015 3:27am
Same style.
For me
Monthly to Daily for long term
Aggressive price action on monthly to determine reversals and Daily to enter using atleat 4 lots. Daily gives me low stop loss. Sometimes I use Weekly to look out for better stop loss levels.
Exits : 1:2, 1:4, third lot is based on Weekly price action(sometimes aggressive and sometimes conservative) and finally monthly exit(only aggressive). However for the fourth lot if it reaches a ratio of 1:10 I close the trade regardless because the recovery can be brutal and I end up losing like 300-400 pips.
Daily to 1 hour chart for swing
and 30 minute to 1 minute for scalping.

For scalping the risk to reward ratio is 1:3, 1:7 and finally reversal on 30 minute(only aggressive)

I trying to improve the Daily to 1 Hour method because the others have been very successful. This one, however, is a pain in the ass and I have to deal with a lot of noise. Maybe in future I'll simply get rid of it.

Best trades are the scalping trades. For some pair R:R is 1:5!
Multi-time frame is analogous to value investing in that you get the best possible trade.
  • #1,020
  • Mar 4, 2015 10:25am Mar 4, 2015 10:25am
Went into this week heavy short Euro. I think the last 6 or so charts I posted were short Euro. I'm not a signal service, but I hope some of my readers got the hint and are doing well on the devaluation of the Euro. If you caught the hint and made some pips, let me know. That would be cool!

Large traders are positioning for tomorrows ECB data, but it was obvious they were positioning for a fall in the Euro a week ago. The best thing about trend trading is we don't really need to know why the market is moving. Often it is moving on inside information that we can't know and the reason why is only obvious after much of the move has occurred.

I can't tell the future and I have no idea how far the Euro will fall, but it was at parity with the USD when issued. That would be down 1000 pips. But, it could turn up in the next 15 minutes. The market can do anything at any time. If you take these multi-lot trend trades in a conservative fashion as I have discussed, funding additional lots with profits as you go and reducing risk all along the way you will have very little risk and the potential reward is great. If it all reverses to stops now I will have made a very tidy sum. I don't expect that, but I'm very well prepared for it.

I did take a lot of EN short off in profit this morning at +225 pips to fund a third lot short on it. Otherwise just riding this Euro down diversified across all pairs except EChf. I still don't trust that manipulated pair.

Going forward, the oil storage in the US is filling up. The latest numbers show a 10 million barrel increase in US storage. What happens when storage is filled up? The oil has to go somewhere and the US can't sell it according to US federal law. Price must drop as the oversupply becomes widespread. The CAD is linked to oil price and If oil drops a lot we can expect the UCad and NCad to soar. It might not happen, but I'm watching for a move. Keep calm and carry on.

G
0 traders viewing now