NFP System - Backtest data provided

Dean,

I've traded it using the P&L line crossing the moving average for signals (2 SMMA on D1, 48 SMMA on M1). By the time the moving average starts to head down, a lot of the move may be over. IMO the market has been consolidating recently, so it's hard to make money trading this basket (unless you're up during London maybe). August was a great month trading this. Maybe after NFP we'll see a definite direction. Trading six currencies has been rough for me. I don't know about others that may trade this. You never know which way the market it really going. The direction for the M1,M5,M30,M60,H4,D1, etc. can added to see overall direction, MACD added may help, but also may hurt. I can't tell you the perfect combination. But one thing is for sure - if we don't have a reliable P&L line to put the moving average or other indicators on, it's all in vain anyways. I too think a script that made our own chart with candles formed using all pairs we want to include would be interesting. My thoughts have been: we can add or subtract pairs and may be able to affect smoothness of the P&L line, and therefore the smoothness of the moving average, to make trading it easier. But one thing is for sure: one either must have patience to wait for the moves and then patience to sit through the move, or we need an EA made. Sometimes there are two big moves a day, but who wants to or can sit 24 hours watching the screen? But again, an EA built on a faulty indicator is not going to be reliable, so first things first: get the indicator fixed or make a script, then the EA.

Don

That formula I quoted is used in USDLFX in Straighthold Trader (custom version of MT4) from lite-forex.org. I never said I would use it. I wouldn't.

it does do averaging. the / in front of the = is saying to divide the total that was previously added together in the indicator, by the number of pairs used. Averaging. You added it for "normalizing", I understand. The real P&L number could just as easily be obtained, instead of taking away that line, to just add another line:
x *= (EUR+JPY+GBP+CHF+CAD+AUD+NZD);

... which would take the normailized/averaged data and mutiply it by the number of pairs used.

If you can, look at it like this:
Say you have 6 pairs, and 3 USD-based pairs gain a pip each (you'd be buying these pairs), and 3 non-usd-based pairs lose a pip each (but you are taking the negative, because you are selling these pairs, so this would be 3 gained pips). You have gained 6 pips. Sentinal would display that as a 1 pip gain (1+1+1+1+1+1=6 & 6/6=1). My version where I removed the line would display 6 pips as the gain (1+1+1+1+1+1=6). That's one way to make P&L. If I just added a line to multiply instead of taking away the line to divide, I would still get 6 (1+1+1+1+1+1=6, 6/6=1, and 1*6=6). That's another way to make P&L, although I just see it as using unnecessary math.

All three versions of the indicator would look the same. The Sentinal Index would have different numbers than the other two, and the other two should have the same numbers theorhetically.

Don

Dean asked how do we know what direction it is going to go. On a high enough timeframe, or using moving averages on either higher or lower timeframes, we can see trend direction. We trade with the trend. I like the "bigger picture", and would like to use M30 or a timeframe around there. Tried to trade M1, and it's just not worth it to me emotionally or physically. I'm ultimately after an EA.

BTW, just looking at the trend this past week:
Starting since last Friday, selling EURUSD, GBPUSD, and AUDUSD, and simultaneously buying USDCHF, USDCAD, and USDJPY, holding for the week, we would be in profit about 1330 pips. Yesterday it was up over 1600 pips. If we could play the intraday charts, not too small timeframe, we could make many more pips. Add good money management, build an EA that can trade the multiple pairs, and you have the best automated business in the world which I have been researching for about 10 months now. I believe God will make it happen. He has already brought me a long way into accomplishing it. I cannot do it alone. Anyone who participates to make it happen, gets a great business for themselves too. So it's a win-win situation. And the more people that use this strategy the better it is for people already using this strategy - as long as we're all using the same timeframe and buying USD and selling USD at the same time. My emotions make it difficult to watch it go against me, so that's why I am ultimately after an EA.

Don

Kevin,

You said that chances were you would have the fixed indicator posted about a week ago. Can you please give an estimated time of arrival on that fixed indicator or script?

Thank you,

Don

Kevin,

Don't know if you've had a chance to try to fix the indicator. If it would be easier for you, maybe a script would be better?

I have a money management equation to add to an EA or indicator or script if needed. If there's anything I can do to help the progress, let me know. Just had an idea to use the indicator in a different way than the original strategy (which you said was fine at the beginning of the thread), so I brought the thread back to life after a few months of no one posting, and about a year since anything real meaningful posted.

I did it to help us all. Have I done something wrong? Sure would like to hear from you. Please don't let this thread die again, when we all could benefit. I have worked too long to see this strategy die like this - it would suck big time. You see, I'm not a pro trader, and it would change my life. That's why I have worked as hard at learning Forex as I have. But my emotions and an indicator needing fixed are obstacles to that goal.

Thanks,

Don

Hey fellas!

Sorry about the delay, Hurricane Ike beat down my home town so I had to do the whole evacuation thing, that on top of my current work put a decent delay on this, but here it is. This is the completely re-written indicator with the iBarShift fix in it, as well as couple of additions. I didn't use any code from the previous indicators (save for the external variables, though I handled them differently). This indicator automatically checks to see if you have data in each of the pairs, and if not, will default to their mini counterparts (if it cannot find values for EURUSD, it will look for EURUSDm automatically). Also, this will find the lowest bar count on each of the 7 pairs to make sure we aren't pulling data from bars that aren't there. For example, if you have 5000 bars in the EUR chart and 2000 bars in the GBP chart on the same timeframe, you do not want to be trying to grab values for the GBP chart on bar 3000 or so, there will be nothing there and you will come back with an incorrect Sentinel value.

Here is the rundown for the external variables:

Weights -- the following values can be set anywhere from 0.1 to infinite. If you don't understand these, please read through the whole thread. Setting the weight to 0 will omit the pair from the calculation.
EUR
JPY
GBP
CHF
CAD
AUD
NZD

AppliedPrice <-- 0 is close, 1 is open, 2 is high, 3 is low. Anything above /below these values is set to close (0) automatically.
MA_Value <-- Moving average value set to 336 by default
MA_Shift <-- Shift for the MA. Set to 0 by default.
MA_Method <-- MA Method (0 is SMA, 1 is EMA, 2 is SMMA, 3 is LWMA). Anything above/below these values is set to SMA (0) automatically.


A quick note or two about this indicator..if you find that your indicator doesn't go very far back into history, that is because you need to download more data for the pair with the lowest number of bars. After you attach Sentinel_Revised, in your Terminal window, click the "Experts" tab. You'll see something that's similar to:

Pair(NZDUSD) bars = #
Pair(AUDUSD) bars = #
Pair(USDCAD) bars = #
...etc...

This will show you how many bars are in each chart. Your indicator will ONLY build values for your LOWEST number here to avoid incorrect values. To download more data, simply open the respective chart and repeatedly tap the PGUP or Home key repeatedly while connected to the data server. Once you obtain more data, simply re-attach the indicator and repeat the process if you feel you still don't have enough data.

Please be kind with your comments here, I coded this in about an hour, and it's definitely almost 5am lol. Look forward to what you guys think about this new revision and please keep a look out for the scripted full chart version soon.

-Kev
Savage Pip Fiend

Kevin,

I've seen that it looks much different than the last indicator internally as well as on the charts. Since I'm working as well as trading, it may take me a little bit of time to get to checking it out.

Don

Kevin,

Forgot to ask: How can I change the indicator (new version) to add all pairs together and not average (so as to use it in a different way than NFP thread has been about)? Or does it not average? It's just hard for me to understand code until I look at it for a bit. Thank you for posting the fixed indicator BTW. Are you going to test it out on the NFP system?

Don

Yeah I started from scratch, and condensed the code with alot of loops to keep it small...also there are a couple of extras in there.

If you want to NOT take the average, simply edit line 168 from:
divNumber++;
to:
// divNumber++;

and change line 172 from:
sentCalc /= divNumber;
to
// sentCalc /= divNumber;

----

Basically just omit the line (by putting // in front of it), and the values you get will just be added, not averaged.

-Kev
Savage Pip Fiend

Hi Kev,

Just a quick thanks so you know your efforts are appreciated. I have not had the time to run the test of the indi, but I look forward to doing that soon. More to come...

Thanks, thanks, thanks.

VBro

I have added 2 indis on top of the sentinal indi window.
1. 50 SMA
2. 18 EMA
3. I have changed the yellow MA to an 89 period

I am using on the 30 min chart.

lookin to trade in the direction indicated by the position of the blue index line ( above or below the 89 EMA ) in addition to 1 or more of the moving averages.
when the 18 ema crosses over the 50 sma, it may give you a heads up of a possible index cross over the 89

Kevin,

First of all, thank you for fixing the indicator. Much appreciated.
I see that you say the default moving average value should be 336, but in the revised indicator, it is set to 100. I made that change to be 336 as default in the indicator, and have posted that as an attached indicator named Sentinal_Revised_Extended.

I don't like using Averaging, so have made this a user Input. It is True by default, meaning each pair's Pip Gain/Loss will be added together, and then divided by the number of pairs whose weight is 1 (number of pairs used). If set to False (the way I plan on using it), each pair's Pip Gain/Loss will be added together. Just like the original indicator, since we are trying to get USD value, pairs that begin with USD are added, and pairs that end with USD have their negative added. Why? Because USD value would be going up when pairs that begin with USD go up, so we would be buying those pairs when they go up. Pairs that end in USD would have to be sold for the USD value to go up when pair is going down, so we add the negative because we aren't buying these pairs as USD is going up, we sell these pairs as USD goes up.

I also don't like having to deal with decimal places, so I have added Input to PipsAsWholeNumbers, which is False by default. If set to True, then as the input notates, we'll see pips as whole numbers. So a gain of 100 pips will be seen as 100, not .0100.

fx__ninja also talked about using 2 more moving averages, so I added those as inputs as well. Use2ndMA and Use3rdMA are set to False by default, but can be set to True and inputs below them set for their Period, Shift, and Method, so that if used, all three MAs can use different periods, shifts, and/or methods.

Let me know what you think. I don't think I messed up the indicator, just added options for users if they wanted to use them.

Also to note: Although I didn't remove the NZD from this attached indicator, even if set to 0, it is looked at to get how many bars are available, so just having NZD there makes the Sentinal and Moving Averages shorter (unless the user also downloads NZD data like you said to using PGUP or Home key).

Hope this helps make the indicator more versatile for more people's needs. I have kept any additions set to False so that the indicator looks like what Kevin last attached by default (with the exception that Moving Average Period is set to 336 instead of 100 per Kevin's post when he attached the Sentinal_Revised indicator).

Hope this helps.

Don

Kev,

Maybe I'm overthinking this...

You use 336 to get the moving average of the previous week right? And it's true there are 336 1/2 hour bars in a 7 day week, but the charts don't have bars for the weekend.

So wouldn't it make sense to figure out how many bars are missing because of the weekend, and then subtract that from 336 to get our M Averages? Or am I missing something?

Just a thought.

VBro

I've found that when I decided to add 'PipsAsWholeNumbers' Input to Sentinal_Revised_Extended, it reads the same on all charts except USDJPY (1745.92 vs. 174592). I know this has something to do with JPY only having two decimal places, but haven't figured out how to fix it yet.

AVERAGING VS. NOT AVERAGING:
As far as averaging vs. not averaging, it looks to be working. If the user wants a "relative" USD index value, use averaging.

If, on the other hand, the user will actually be buying the usd-based pairs and selling the non-usd-based pairs when the Sentinal line is going up, and wants to know how many total pips would be gained from one point on the sentinal line to a higher point on the sentinal line if "buying usd", and want the "absolute" pip gain, don't use averaging. If "selling USD", when Sentinal line going down, buy non-usd-based pairs while selling usd-based pairs.

NOW WE TALKIN' 'BOUT SOME PIPS! If we could use whole numbers and not decimals, it brings this reality out better when looking at the numbers and subtracting one point on line from another to get TOTAL pips gained using/trading all the pairs together. I will try to fix it to read same on all pairs' charts, including USDJPY, or maybe Kevin can look at it and see if he can fix it. Either way, this is looking good.

Don

Had the Sentinal Revised Extended on EURUSD M1 chart, and it looks like between 09.22.08 1:54 to 09.22.08 2:19, that if "selling USD" (buying EURUSD,GBPUSD,AUDUSD and selling USDCHF,USDCAD,USDJPY) because it crossed below moving average, could have made almost 1000 pips in less than a half hour? (It's moved back up a good little bit, so could have made some now "buying USD")

1000 pips in that amount of time. Is this number right? If so, that's SO COOL!

LET'S CHECK (numbers are approx.):
EURUSD buy 1.4438, sell 1.4489 +51
GBPUSD buy 1.8303, sell 1.8329 +26
AUDUSD buy 0.8292, sell 0.8303 +11
USDCHF sell 1.1052, buy back 1.1009 +43
USDCAD sell 1.0490, buy back 1.0491 +1
USDJPY sell 107.00, buy back 106.43 +57

So when we add those together, we get 189 pips in less than a half hour, not 1000. So numbers are not correct, but drawn Sentinal appearance remains unchanged, so can use revised or revised extended to do these types of trades. The problem is how there's only two decimals after USDJPY vs 4 decimals for the others. It looks to affect both sentinal revised and sentinal revised extended indicators, which needs checked again.

Don

Just wanted others to note I was using FXDD when I told the times in the last email. If using MT4 download from MetaQuotes, it would be 09.21.08 22:54 to 09.21.08 23:19. It seems to me that FXDD's download keeps up with changes faster, so I use it, with MetaQuotes MT4 as a backup if I need it.

Don

disbellj,

Just FYI, the reason you are getting decimal places is that you are using the constant "Point" to divide by. Constants depend on what chart you have the indicator/EA attached to. When you attach it to any chart besides a JPY based chart, you are dividing your entire number basically by the 2 constant instead of the 4 constant...and 4 being the max number of decimal places in our USD based pairs, you end up with a whole number. By attaching your indicator to a Yen chart, you end up dividing by the 2 constant, and you get only 2 decimal places removed. You can solve this by simply dividing it by 1000 instead of leaving it to the Point constant.

-Kev
Savage Pip Fiend

Kevin,

I remember I had to use 100 for USDJPY and 10000 for all others when I made my indicator (that didn't update correctly), to get the correct pip gains for all pairs. Trying 1000 for all places gives me results that match on all charts, but the range totals not even one pip, so I know this is not what I'm after. So I'll try to figure it out (probably will have to break USDJPY out from the rest and calculate it differently).

Thank you,

Don

BTW, I'm using Average set to False, and PipsAsWholeNumbers set to True.

Don

When using PipsAsWholeNumbers, I'm having to multiply USDJPY by 100, and multiply everything else by 10000, so no division. Once I get this figured out how to fix it in the code to read correctly when using multiple pairs, I'll post an update.

Don