AshFX Daily (EA - MTaboneWeb) | Page 3 | Forex Factory

Options

Go to
Permalinks
Print Thread

Similar Threads

AshFX Daily 2,650 replies

DailyHighLow (EA - MTaboneWeb) 103 replies

"The Pip Nailer" (EA - MTaboneWeb) 74 replies

GrabMyOrders (EA - MTaboneWeb) 20 replies

TossUp (EA - MTaboneWeb) 9 replies

Attachments: AshFX Daily (EA - MTaboneWeb)
Exit Attachments

AshFX Daily (EA - MTaboneWeb)

  • #41
  • Jul 20, 2008 12:19am Jul 20, 2008 12:19am
does your EA calculate the S/R too?
  • #42
  • Jul 20, 2008 8:39am Jul 20, 2008 8:39am
Quoting silvercoin
will EA auto adjust the SL and TP?

At this point the EA sets everything based off the original system. I have plans to implement some additions like trailing stop and even a trailing stop that's based on the PSAR value although it doesn't do either right now.
  • #43
  • Jul 20, 2008 8:42am Jul 20, 2008 8:42am
Quoting moshnjack
does your EA calculate the S/R too?

I'm still looking for a reliable indicator that can be used for S/R. I was using "Support and Resistance (Barry)" manually but it doesn't seem to do very well. In some cases it went well beyond the S/R and made profit and then sometimes not so it was kind of a toss up.

Any suggestions? Maybe one that does well in a daily time frame?
  • #44
  • Jul 21, 2008 12:25pm Jul 21, 2008 12:25pm
I added a Settings section to the first post to explain the options.
  • #45
  • Jul 21, 2008 2:13pm Jul 21, 2008 2:13pm
Please check post 1.
  • #46
  • Jul 21, 2008 2:14pm Jul 21, 2008 2:14pm
The two trades above were opened yesterday but closed today by the EA without hitting their targets or stop loss. Closed out shortly after the change of day.

Am I doing something wrong when implementing the EA?
James The Beekeeper
  • #47
  • Jul 21, 2008 2:16pm Jul 21, 2008 2:16pm
Quoting edeyj999
The two trades above were opened yesterday but closed today by the EA without hitting their targets or stop loss. Closed out shortly after the change of day.

Am I doing something wrong when implementing the EA?

I believe they hit their exit criteria according to the system (AO changed colors).
  • #48
  • Jul 21, 2008 2:54pm Jul 21, 2008 2:54pm
Quoting mtaboneweb
I believe they hit their exit criteria according to the system (AO changed colors).
Don't think so. Yes, they changed to green but this is not an exit criteria for a long.
James The Beekeeper
  • #49
  • Jul 21, 2008 3:19pm Jul 21, 2008 3:19pm
Quoting edeyj999
Don't think so. Yes, they changed to green but this is not an exit criteria for a long.
Actually it is. See the screen shot of the pdf. You can also see the portion explaining that for a second strength trade the AO doesn't have to be compliant and that it will usually change colors but in this case did not. I didn't develop the system, only the EA to trade it according to the rules. The EA did what it was supposed to do according to the rules.

This may require some tweaking for special scenarios since it's mechanical rather than manual.
Attached Image (click to enlarge)
Click to Enlarge Name: rules.gif Size: 42 KB
  • #50
  • Jul 21, 2008 5:24pm Jul 21, 2008 5:24pm
Quoting mtaboneweb
Actually it is. See the screen shot of the pdf. You can also see the portion explaining that for a second strength trade the AO doesn't have to be compliant and that it will usually change colors but in this case did not. I didn't develop the system, only the EA to trade it according to the rules. The EA did what it was supposed to do according to the rules.

This may require some tweaking for special scenarios since it's mechanical rather than manual.
I would seriously question whether the AO changing to green on a long trade is a criterion for exiting the trade.

I took a look at the screenshot as you recommended, but that only reinforces my opinion. Regarding exiting a trade, it states "So if you're long, you would exit the second half of your position when the AO eventually turns red." In this case, it turned green, which further validated the long position.

The screenshot also notes, in regard to SS trades, "when taking a long, if the AO isn't compliant, one can still take the trade; it will usually turn green within the next couple of candles." That's exactly what happened, so why would that be a basis for exiting the position?
Greetings from Guanajuato, Mexico
  • #51
  • Jul 21, 2008 8:46pm Jul 21, 2008 8:46pm
Quoting yen44x
I would seriously question whether the AO changing to green on a long trade is a criterion for exiting the trade.

I took a look at the screenshot as you recommended, but that only reinforces my opinion. Regarding exiting a trade, it states "So if you're long, you would exit the second half of your position when the AO eventually turns red." In this case, it turned green, which further validated the long position.

The screenshot also notes, in regard to SS trades, "when taking a long, if the AO isn't compliant, one can still take the trade; it will usually turn green within the next couple of candles." That's exactly what happened, so why would that be a basis for exiting the position?

It changed to green on the current bar. The EA and the system is based off of completed bars. It should be clear looking at the chart that the previous bar (not the bar where the trade was closed) was even lower than the bar before that (which was the signal bar). On the very next bar it placed the long orders at the open of the bar and when it rolled over to today's bar it hadn't increased. The bar where the buy was initiated was lower the the bar where the trade was signaled. forexmoments has said repeatedly that he works on completed bars. It does not matter that the current bar is turning green because the previous closed bar was the signal for it to close.

Like I've said before, I did not develop the system. I only developed the EA to trade the system precisely by the rules given by forexmoments. In this particular case it could end up being unfortunate that the trades closed early. I totally agree at this very moment with you but if it had continued down or does turn around before the end of this bar and be even lower then you would be upset that the trade lost even more money hitting a stop loss possibly. As far as I'm concerned the EA did exactly what forexmoments laid down as the rules.

People who trade this system manually or with an EA need to ignore the current bar and look bar at the previous closed bars. That is where all the decisions are made.
Attached Image (click to enlarge)
Click to Enlarge Name: ashfx_eurjpy.gif Size: 29 KB
  • #52
  • Edited 10:00pm Jul 21, 2008 9:15pm | Edited 10:00pm
Let me clarify, I have no ax to grind at all. I don't use the EA yet, but appreciate all the hard work that went into it. In fact, I did not take this trade manually either because I felt the long signal came too close to past resistance. In this specific instance, I felt compelled to post because I thought edeyj999 made a valid point, and it would give you an opportunity to explain how the EA works just so we all could understand it a little better. Let me go through some of the points that you made and perhaps you can clarify my understanding of how you interpret the AshFX rules.

Quoting mtaboneweb
It changed to green on the current bar. The EA and the system is based off of completed bars. It should be clear looking at the chart that the previous bar (not the bar where the trade was closed) was even lower than the bar before that (which was the signal bar).
That's absolutely correct. As I stated previously, the green candle appeared exactly two bars after the signal candle, which Ash noted often happens on SS trades. In regard to the height of the bar, I wasn't aware that was taken into account in deciding whether a trade would end. Are you saying the EA detected that and closed the trade for that reason?

Quote
On the very next bar it placed the long orders at the open of the bar and when it rolled over to today's bar it hadn't increased. The bar where the buy was initiated was lower the the bar where the trade was signaled. forexmoments has said repeatedly that he works on completed bars. It does not matter that the current bar is turning green because the previous closed bar was the signal for it to close.
Are you saying the trade was closed by the EA due to the fact that the bar was lower than the previous one? Again, I wasn't aware that the height of the bar mattered, but if you programmed that into the EA, it's good to know.

Quote
Like I've said before, I did not develop the system. I only developed the EA to trade the system precisely by the rules given by forexmoments. In this particular case it could end up being unfortunate that the trades closed early. I totally agree at this very moment with you but if it had continued down or does turn around before the end of this bar and be even lower then you would be upset that the trade lost even more money hitting a stop loss possibly. As far as I'm concerned the EA did exactly what forexmoments laid down as the rules.
I certainly understand you did not develop the system, and I'm not trying to give you flak for being kind enough to develop an EA for the system. However, I would still respectfully disagree that this trade closed "precisely" by AshFX rules, unless I'm totally misunderstanding something about the system. The trade never reached the stop loss. If you're saying it closed because of the red AO bar on 7/21, doesn't that actually conflict with AshFX rules for SS trades, since the rules mention that the AO will "usually turn green with the next couple of candles", which is exactly what happened. I understand it's impossible to program every possible scenario into an EA, and perhaps this is one of those scenarios, but I just don't see how closing this specific trade conforms "exactly" to AshFX rules.

Quote
People who trade this system manually or with an EA need to ignore the current bar and look bar at the previous closed bars. That is where all the decisions are made.
I agree.

I've made it a practice not to ever criticize the trade decisions of others, or to denigrate the work of a well-intended EA programmer, and, again, that's certainly not my intent here. It could be that I'm still misunderstanding what you're saying.
Greetings from Guanajuato, Mexico
  • #53
  • Jul 21, 2008 10:50pm Jul 21, 2008 10:50pm
Quoting yen44x
In regard to the height of the bar, I wasn't aware that was taken into account in deciding whether a trade would end. Are you saying the EA detected that and closed the trade for that reason?

The color of the bar is determined by how it compares to the previous bar. If it's less than then it's red and if it goes higher then green. Each progressively larger/higher bar will continue to be green until the first bar that is less and then that bar will be red. In the EA I don't see Red/Green I see values higher/lower (less than / greater than).

I believe there would have to be something custom to be able to accommodate a second strength trade if the AO doesn't comply when the trade is initiated. On one hand the AO is specifically used as an exit and then dismissed for a period of time in the case of a second strength trade. The EA would have to be taught the difference and yet still know when it actually needed to use the AO as a signal to close the trade(s). Perhaps a variable for number of bars it's excused before it uses the AO as an exit possibility? You see what I mean? Even if both trades are open and the AO changes under the right circumstances then I feel both should be closed even if the first hasn't hit the take profit. It may not hit the take profit thus leaving it to double back and hit the stop loss eventually.

In this case an extra bar of forgiveness might have done the trick but how much slack do you give these things? It also might have been a bad thing as well. In a manual world then the system would have human intervention but the EA is purely technical only doing what it was told to do.

Suggestions?
  • #54
  • Edited 11:38pm Jul 21, 2008 11:17pm | Edited 11:38pm
Quoting mtaboneweb
The color of the bar is determined by how it compares to the previous bar. If it's less than then it's red and if it goes higher then green. Each progressively larger/higher bar will continue to be green until the first bar that is less and then that bar will be red. In the EA I don't see Red/Green I see values higher/lower (less than / greater than).
That certainly makes sense, and I see now why you referred to the height of the bar.

Quote
I believe there would have to be something custom to be able to accommodate a second strength trade if the AO doesn't comply when the trade is initiated. On one hand the AO is specifically used as an exit and then dismissed for a period of time in the case of a second strength trade. The EA would have to be taught the difference and yet still know when it actually needed to use the AO as a signal to close the trade(s).
Now I know why the word "conundrum" was invented . . . just to describe this situation, because I'm not sure there's a ready answer for that problem. It would be interesting to have Ash comment on how often SS trades includes a red AO bar. If it's relatively rare, perhaps the EA could "ignore" those trades. To my admittedly untrained eye, that seems easier than developing an EA that could somehow distinguish when to continue or exit a trade based on a rule that varies based on the type of trade (Standard vs. SS).

Quote
Perhaps a variable for number of bars it's excused before it uses the AO as an exit possibility? You see what I mean? Even if both trades are open and the AO changes under the right circumstances then I feel both should be closed even if the first hasn't hit the take profit. It may not hit the take profit thus leaving it to double back and hit the stop loss eventually.
I see exactly what you mean, and I don't see that there's an easy solution to those issues.

Quote
In this case an extra bar of forgiveness might have done the trick but how much slack do you give these things? It also might have been a bad thing as well. In a manual world then the system would have human intervention but the EA is purely technical only doing what it was told to do.
Understood, and again, no instant "fix" comes to mind on this specific problem.

Quote
Suggestions?
During the height of popularity for the OzFX thread, I was asked by one of the EA developers on that thread to test the EA prior to its publication and give honest, candid feedback on any flaws, "bugs", or failure to conform to system rules that I observed, and we exchanged dozens of e-mails in the process. As a result, I'm used to doing that, and again, meant no disrespect toward your efforts. I appreciate that developing an EA to everyone's satisfaction is almost impossible and often a thankless task.

One of the reasons for noting the issues I saw was to see if it solicited some suggestions from users of the EA or just interested observers. Sometimes the solution is easier than it would seem at first glance, so perhaps some traders would like to offer their input.
Greetings from Guanajuato, Mexico
  • #55
  • Jul 22, 2008 12:27am Jul 22, 2008 12:27am
this EA will not attach to my charts. any ideas on what the problems might be?
  • #56
  • Edited 1:24pm Jul 22, 2008 7:42am | Edited 1:24pm
Well, after a good night's sleep, I do have a couple of suggestions regarding how the EA might handle SS trades where the AO bar does not conform i.e. still green on a Short trade, still red on a Long trade.

I took a look at the 4 major currency pairs over the past 6 months. Over that span, I found 30 trades that would be defined as AshFX SS trades. Of those, the vast majority, 26, conformed completely (in other words, the AO bar conformed with the rest of the trade signal). In only 4 cases was there a situation similar to what was encountered on the eur/jpy trade we were discussing.

If that same percentage is somewhat similar for all the currency pairs, would it be possible for the EA to do one of these two suggestions?:

1. Ignore a trade altogether if the AO is not compliant. I realize that does not conform completely with AshFX rules, but it also prevents trying to figure out how to program the EA to exit per AshFX rules in this type of trade scenario. Of the 4 trades I found where the AO was not compliant, some were losing trades and none turned into large winners, so I'm not sure that eliminating them from the trading system would change the results much at all.

On average, this type of trade situation only came up once every 6 months per currency pair on the 4 pairs I tested, so eliminating the "non-conforming" trade would not seem to affect overall performance much at all.

2. Open the trade only when the AO is compliant with the rest of the trade signal. As the rules state, that usually happens within a couple of bars anyway.

Neither of these suggestions is perfect, since they do not match the AshFX rules exactly, but at this point, the EA exit strategy for this type of trade is not perfect either. All in all, it might prove to be a more manageable way for the EA to handle the trade and would eliminate a repeat of the EUR/JPY exit situation.
Greetings from Guanajuato, Mexico
  • #57
  • Jul 22, 2008 5:11pm Jul 22, 2008 5:11pm
Quoting yen44x
Well, after a good night's sleep, I do have a couple of suggestions regarding how the EA might handle SS trades where the AO bar does not conform i.e. still green on a Short trade, still red on a Long trade.

I took a look at the 4 major currency pairs over the past 6 months. Over that span, I found 30 trades that would be defined as AshFX SS trades. Of those, the vast majority, 26, conformed completely (in other words, the AO bar conformed with the rest of the trade signal). In only 4 cases was there a situation similar to what was encountered on the eur/jpy trade we were discussing.

If that same percentage is somewhat similar for all the currency pairs, would it be possible for the EA to do one of these two suggestions?:

1. Ignore a trade altogether if the AO is not compliant. I realize that does not conform completely with AshFX rules, but it also prevents trying to figure out how to program the EA to exit per AshFX rules in this type of trade scenario. Of the 4 trades I found where the AO was not compliant, some were losing trades and none turned into large winners, so I'm not sure that eliminating them from the trading system would change the results much at all.

On average, this type of trade situation only came up once every 6 months per currency pair on the 4 pairs I tested, so eliminating the "non-conforming" trade would not seem to affect overall performance much at all.

2. Open the trade only when the AO is compliant with the rest of the trade signal. As the rules state, that usually happens within a couple of bars anyway.

Neither of these suggestions is perfect, since they do not match the AshFX rules exactly, but at this point, the EA exit strategy for this type of trade is not perfect either. All in all, it might prove to be a more manageable way for the EA to handle the trade and would eliminate a repeat of the EUR/JPY exit situation.

I would have to say suggestion 1 is the easiest and maybe the best as well. If suggestion 2 were implemented then the EA would need to look back further than it does now looking for a trade that might have been close but AO was compliant. In doing that it might lead to bugs allowing trades to be placed when they weren't valid to begin with.

I really do appreciate any and all suggestions. It's just been a little crazy maintaining 4 of my own threads and EAs for each one plus bug testing, etc. I spent the entire day today bug testing my WeeklyOpen EA. It was very frustrating. Some times it seems you have to over do everything when writing an EA just to make sure there's no way for the broker to mess it up and then sometimes it's just a bug that stays hidden for a good long while.

Thanks for your help/support.
  • #58
  • Jul 22, 2008 6:58pm Jul 22, 2008 6:58pm
For a trade to be compliant, doesn't a long trade need AO to go short and a short trade look for a long AO?

Red AO is simply current AO less than previous AO
Green AO is simply current AO greater than previous AO

If we opened long, we're looking for a change from green to red, so--assuming 1 is the current bar--we're looking for 1 < 2 > 3.

If we opened short, we're looking for a change from red to green, so--assuming 1 is the current bar--we're looking for 1 > 2 < 3.

I don't see why this can't be implemented...
  • #59
  • Jul 23, 2008 2:00am Jul 23, 2008 2:00am
Folks,
I have just come back to this thread after a couple of days away and see that my original posts have generated some discussion!

Just to put my cards on the table and as a preamble to my comments below I would like to state a few things:

1. Anything I say is not a criticism but is meant to be constructive.
2. The way the EA exits trades will not really affect me because I only intend to use the EA to confirm my interpretation of possible trade entries (although I had hoped it might be useful to determine if ASHFx weekly would be profitable just by taking the trade signals and exiting per the rules with no human intervention).
3. I am a system developer and know how difficult it can be to create something from a user specification. Ash has done a good job with his specification but in my opinion there are some areas that need further clarification. I think this is borne out by the fact that we disagree with the exit rules for these trades.
4. My view of the system is exactly that of YEN44X for these trades.

The points I would like to make are as follows:

1. When I read the rules, it suggests to me that only the second half of a trade should look for a change of colour of AO. And then, for a long, the AO would turn from green to red, and vice versa for a short trade for the trade to be closed (not just a change of colour regardless). Now, I do have a problem with this in that there have been trades recently where the first part of the trade has not hit SL or TP before the AO changes colour. Ash's specification does not state what to do in these circumstances. Leaving those trades to run invariably means that they (both halves) hit SL with large losses (due to the SL being set at the prior swing high or low). As a result, I have been monitoring the system from the start of July and, for closed trades, the system is in a large loss. I think further clarification is needed for this scenario from Ash.

2. You mention only taking trading decisions on completed candles. I don't think this is correct. The decision for entry is only ever taken on a closed candle but the exit decision is taken on the current candle. SL or TP will only ever happen during a trading session so will be for a current candle. The only other exit criteria is when the current bar on the AO changes colour - again, this can only happen on the current open candle.

3. My final point is regarding trade entry. These trades entered when both the AO conflicted and the stochs were above/below their ideal. The way I read the spec these are independent trade factors. There is nothing in the spec that says the trade is valid if both these factors occur. Ash would need to confirm that the combination of the two still makes a valid trade.

So, that's my take on things. I would be happy to participate in any discussions you may have regarding these or other rules of the system. I am also happy to try to contact Ash for clarification if you wish. I am also mindful that the EA is yours and you must decide how you want to play it. Those of us that decide to use it do so in the knowledge that any trade we place is at our risk not that of the EA.

Sorry for the length of the post but I thought I may be able to help. Thanks for your efforts so far.
James The Beekeeper
  • #60
  • Jul 23, 2008 8:26am Jul 23, 2008 8:26am
Quoting edeyj999
1. When I read the rules, it suggests to me that only the second half of a trade should look for a change of colour of AO. And then, for a long, the AO would turn from green to red, and vice versa for a short trade for the trade to be closed (not just a change of colour regardless). Now, I do have a problem with this in that there have been trades recently where the first part of the trade has not hit SL or TP before the AO changes colour. Ash's specification does not state what to do in these circumstances. Leaving those trades to run invariably means that they (both halves) hit SL with large losses (due to the SL being set at the prior swing high or low). As a result, I have been monitoring the system from the start of July and, for closed trades, the system is in a large loss. I think further clarification is needed for this scenario from Ash.
I just found this.

Ash states that AO is not considered until first TP is hit and then anytime after that could close the 2nd trade. If both are open then regardless of AO color the trades both remain active until 1st TP is hit or the both hit the SL. Of course this is not explained thoroughly in the pdf but it's news to me now and something I think can be implemented relatively easily. Do you agree?

Quoting edeyj999
3. My final point is regarding trade entry. These trades entered when both the AO conflicted and the stochs were above/below their ideal. The way I read the spec these are independent trade factors. There is nothing in the spec that says the trade is valid if both these factors occur. Ash would need to confirm that the combination of the two still makes a valid trade.
I agree, if you would like to get further verification that this is the case then please let me know.
0 traders viewing now