(binned per thread starter’s request) MM (Money Maker) Detective | Page 203 | Forex Factory

Options

Go to
Permalinks
Print Thread

Similar Threads

Money management: Capital needed per 0.07 USD per microlot 0 replies

The Volume Detective 411 replies

Attachments: (binned per thread starter’s request) MM (Money Maker) Detective
Exit Attachments

(binned per thread starter’s request) MM (Money Maker) Detective

  • #4,041
  • Jul 1, 2019 12:43am Jul 1, 2019 12:43am
Quoting JaxPacific
{quote} For sure. Same conclusion, for the most part. In my experience, if you find a good system, then start over optimizing and retrofitting, it can take a good system that works 70% of the time.. and turn it into a system that works 90% of the time but only until the market behavior changes. Most people who... I can't believe I'm saying this.. I vowed never to say this on forex factory because of my loath... forget it , can't do it, I tried.. I'll just say people looking for the HG.. they geek out over code and forget or don't know why the trades...
Quote
study the indicators and find me some bad/failed trades on any pairs except EUR USD, USD JPY or USD CHF. I'll study those.
@JaxPacific -- excellent. I'm interested to learn more from you and what you look for.

And I'm curious -- you don't like EURUSD, USDJPY, USDCHF -- why? What's different about them?
  • #4,042
  • Jul 1, 2019 12:46am Jul 1, 2019 12:46am
Oh they're fine. I mean, the Eur is probably one of the most manipulated pairs, but I still trade it. I just meant that I am already scouting those 3 myself.
  • #4,043
  • Jul 1, 2019 12:52am Jul 1, 2019 12:52am
Quoting bishopdotun
{quote} _______________________________ R4M - clever dog I thought the same, but as the problem is the VPS machine itself (or at least I think) having limited resources, chopping it up in different MT4 installs (sucking up the same RAM etc) - would probably give same results....anyway, there is only one way to find out! My wet dream is still a simple dashboard firing off alerts (closed candle for less CPU usage) - when there is a set-up. Are we 100% sure the BDADR does not repaint? I use it on 1H chart, but still set to 10080 in the TF inputs....
We wont get tired of that repainting problem. You may notice minor movements here and there but they are not significant. The entry points mostly remain the same. Thats why I chose weekly and also monthly. Additionally, we are less bothered because price has already overshot its boundaries (>100% ADR for the period relative to centred moving avg) at the moment the trade was taken. So we expect price to return to its fair value soon. And we keep doing that each time price gets away from its fair value. Also from historical perspective, price almost always touch the middle line at least once a week giving us opportunities to exit in profit or loss. We will scale in to the trades by taking microlots, spacing the entry points and setting our stop loss at a further x% of ADR. Combining these give us a clear edge. At least visual backtest shows that. Overtime we should make more than we expect to loose. @R4M: This is why the automation is necessary. I missed this while asleep. I have gone back to the previous version. It is lighter. The new one is killing the platform. {image}[/quote]

___________

Exactly how I play it (if I see the signals in time) taking the trades when a signal and scaling into profitable trades, took 2 set-ups so far..

I prefer to take the initial trade manually (to scout for other conditions etc) - and then let some trademanger EA do the rest.

PC
Attached Image(s) (click to enlarge)
Click to Enlarge Name: UC long.png Size: 54 KB
Click to Enlarge Name: EA long.png Size: 40 KB
Less is more...
  • #4,044
  • Jul 1, 2019 1:14am Jul 1, 2019 1:14am
Quoting bishopdotun
{quote} _______________________________ R4M - clever dog I thought the same, but as the problem is the VPS machine itself (or at least I think) having limited resources, chopping it up in different MT4 installs (sucking up the same RAM etc) - would probably give same results....anyway, there is only one way to find out! My wet dream is still a simple dashboard firing off alerts (closed candle for less CPU usage) - when there is a set-up. Are we 100% sure the BDADR does not repaint? I use it on 1H chart, but still set to 10080 in the TF inputs....
We wont get tired of that repainting problem. You may notice minor movements here and there but they are not significant. The entry points mostly remain the same. Thats why I chose weekly and also monthly. Additionally, we are less bothered because price has already overshot its boundaries (>100% ADR for the period relative to centred moving avg) at the moment the trade was taken. So we expect price to return to its fair value soon. And we keep doing that each time price gets away from its fair value. Also from historical perspective, price almost always touch the middle line at least once a week giving us opportunities to exit in profit or loss. We will scale in to the trades by taking microlots, spacing the entry points and setting our stop loss at a further x% of ADR. Combining these give us a clear edge. At least visual backtest shows that. Overtime we should make more than we expect to loose. @R4M: This is why the automation is necessary. I missed this while asleep. I have gone back to the previous version. It is lighter. The new one is killing the platform. {image} USDCAD: According to rules, UsdCad long trades should exit now.[/quote]

@bishopdotun -- excellent explanation -- thank you.

Quote
I have gone back to the previous version. It is lighter. The new one is killing the platform.
Okay. The two main differences are the period separator and communicating with the proxy EA. But before reverting the code, can you describe a scenario where you've noticed a change in performance. The reason I ask is then that would give me a pseudo-metric to compare different versions. I'm also going to add an elapsed timer diagnostic to the indicator's main function that will compute (and display) how long it takes to process each tick -- and in that way we could better monitor not only its performance on my machine but on everyone else's, as well.
  • #4,045
  • Jul 1, 2019 1:18am Jul 1, 2019 1:18am
Quoting robots4me
{quote} We wont get tired of that repainting problem. You may notice minor movements here and there but they are not significant. The entry points mostly remain the same. Thats why I chose weekly and also monthly. Additionally, we are less bothered because price has already overshot its boundaries (>100% ADR for the period relative to centred moving avg) at the moment the trade was taken. So we expect price to return to its fair value soon. And we keep doing that each time price gets away from its fair value. Also from historical perspective, price...
@bishopdotun -- excellent explanation -- thank you. {quote} Okay. The two main differences are the period separator and communicating with the proxy EA. But before reverting the code, can you describe a scenario where you've noticed a change in performance. The reason I ask is then that would give me a pseudo-metric to compare different versions. I'm also going to add an elapsed timer diagnostic to the indicator's main function that will compute (and display) how long it takes to process each tick -- and in that way we could better monitor not only its performance on my machine but on everyone else's, as well.[/quote]

_________

Couldn't we just make the indicator check every X candle or X time (ie every 30 M)?

I know the exit touch would be late then, but we would survive that...

PC
Less is more...
  • #4,046
  • Jul 1, 2019 1:26am Jul 1, 2019 1:26am
Quoting bishopdotun
{quote} _______________________________ R4M - clever dog I thought the same, but as the problem is the VPS machine itself (or at least I think) having limited resources, chopping it up in different MT4 installs (sucking up the same RAM etc) - would probably give same results....anyway, there is only one way to find out! My wet dream is still a simple dashboard firing off alerts (closed candle for less CPU usage) - when there is a set-up. Are we 100% sure the BDADR does not repaint? I use it on 1H chart, but still set to 10080 in the TF inputs....
We wont get tired of that repainting problem. You may notice minor movements here and there but they are not significant. The entry points mostly remain the same. Thats why I chose weekly and also monthly. Additionally, we are less bothered because price has already overshot its boundaries (>100% ADR for the period relative to centred moving avg) at the moment the trade was taken. So we expect price to return to its fair value soon. And we keep doing that each time price gets away from its fair value. Also from historical perspective, price almost always touch the middle line at least once a week giving us opportunities to exit in profit or loss. We will scale in to the trades by taking microlots, spacing the entry points and setting our stop loss at a further x% of ADR. Combining these give us a clear edge. At least visual backtest shows that. Overtime we should make more than we expect to loose. @R4M: This is why the automation is necessary. I missed this while asleep. I have gone back to the previous version. It is lighter. The new one is killing the platform. {image} USDCAD: According to rules, UsdCad long trades should exit now.[/quote]

Quote
UsdCad long trades should exit now.
@bishopdotun -- I was just about to respond to your screen shot but then it morphed from USDCAD to NZDJPY

In the previous version I think I was checking for each new bar -- in which case a High crossing the midline would not have generated an Exit signal -- only a Close would have. But in the version I'm currently working then I do allow for each tick to be processed and a High should also generate an Exit signal.
  • #4,047
  • Jul 1, 2019 1:39am Jul 1, 2019 1:39am
Quoting pipcruiser
{quote} @bishopdotun -- excellent explanation -- thank you. {quote} Okay. The two main differences are the period separator and communicating with the proxy EA. But before reverting the code, can you describe a scenario where you've noticed a change in performance. The reason I ask is then that would give me a pseudo-metric to compare different versions. I'm also going to add an elapsed timer diagnostic to the indicator's main function that will compute (and display) how long it takes to process each tick -- and in that way we could better monitor...
_________ Couldn't we just make the indicator check every X candle or X time (ie every 30 M)? I know the exit touch would be late then, but we would survive that... PC[/quote]

@pipcruiser -- what do you mean by "every X candle"? We make trading decisions based on the tick price -- now.
  • #4,048
  • Jul 1, 2019 2:04am Jul 1, 2019 2:04am
Another missed trade (while asleep)

.....and it was a sling shot.

Attached Image (click to enlarge)
Click to Enlarge Name: Missed.png Size: 1.0 MB


This one is a clear example of some people "showing their hand". Whoever they are, I don't bother.

Price gets away from fair value(midline), i shoot. And I want to see all pairs in just one dashboard window. The signal appears, the bot shoots.

Price gets to the midline, the bot cuts the trade in profit or loss.

Repeat....Repeat.
  • #4,049
  • Jul 1, 2019 2:21am Jul 1, 2019 2:21am
Quoting bishopdotun
Another missed trade (while asleep) .....and it was a sling shot. {image} This one is a clear example of some people "showing their hand". Whoever they are, I don't bother. Price gets away from fair value(midline), i shoot. And I want to see all pairs in just one dashboard window. The signal appears, the bot shoots. Price gets to the midline, the bot cuts the trade in profit or loss. Repeat....Repeat.
@bishopdotun -- nice.

A question about the dashboard. Is it good enough for the dashboard to simply say BUY or SELL or EXIT, along with the server and local times?
  • #4,050
  • Jul 1, 2019 2:26am Jul 1, 2019 2:26am
Scaling the trades: Would you click buy or sell each time?

I have circled all the signals on H1 on AudJpy in the past few weeks. All winners.

Attached Image (click to enlarge)
Click to Enlarge Name: Missed.png Size: 1.1 MB


So easy in hindsight and we know the edge is on our side. The odds of winning is great.
We know the fair value has been overshot and price will find its way back. But we don't know when.
So we don't want to overcommit in case price doesn't revisit the fair value soon enough. So we need to cover the ass.

Here is what I think would work.

- You have $10k in the account
- You take the trades at 0.01 (do we have brokers who offer 0.005?)
- Take the first signal
- Space the trades at 10%-20% of ADR (only signals that meet this requirements after the first signal will be taken)
- Set stop loss for each trade to 150%-200% ADR (There are very very few occasions where this was breached across all pairs)
- Set the bot with the criteria above and avoid emotions. Even if the market would crash, stop loss would cover the ass
- Make sure your broker is not a criminal.

Off to my day job now.

Ciao.
  • #4,051
  • Jul 1, 2019 2:51am Jul 1, 2019 2:51am
Quoting robots4me
{quote} @bishopdotun -- nice. A question about the dashboard. Is it good enough for the dashboard to simply say BUY or SELL or EXIT, along with the server and local times?
I sent a PM.

Attached Image (click to enlarge)
Click to Enlarge Name: Sample.png Size: 700 KB
  • #4,052
  • Jul 1, 2019 2:55am Jul 1, 2019 2:55am
Quoting bishopdotun
Exit & Entry The white rectangle mark is both an entry and exit {image} Remember: - If we are on a sell and price touches the lower band, we exit and vice versa. - Touch means price poking above the line by at least 1pip.
@R4M: I guess you missed this post.

Just reminding you. There would be cases where exit is opposite band.
  • #4,053
  • Edited 3:23am Jul 1, 2019 3:08am | Edited 3:23am
Quoting bishopdotun
{quote} I sent a PM. {image}
@bishopdotun -- WOW. I'll have it completed by tomorrow morning, at the latest.

Okay, this gives me an idea of what you had in mind. Let me first create something simple, yet functional -- more like an "Alert" dashboard.

And then if this BD strategy continues to get high grades from people then I would consider creating something more sophisticated.

This is all too cool. @bishopdotun -- thank you.
  • #4,054
  • Jul 1, 2019 3:15am Jul 1, 2019 3:15am
Quoting bishopdotun
Another missed trade (while asleep) .....and it was a sling shot. {image} This one is a clear example of some people "showing their hand". Whoever they are, I don't bother. Price gets away from fair value(midline), i shoot. And I want to see all pairs in just one dashboard window. The signal appears, the bot shoots. Price gets to the midline, the bot cuts the trade in profit or loss. Repeat....Repeat.

Nobody is showing his hand other than Centered Moving Averages crossing in the middle of Price Action Swings.

A pure arithmetical calculation

A very simple algorithm.
Attached Image(s) (click to enlarge)
Click to Enlarge Name: AUDJPY MOVING AVERAGES CROSSINGS 1 JULY 2019.png Size: 1.1 MB
Click to Enlarge Name: BISHOP SHOWING THEIR HAND.png Size: 1.0 MB
  • #4,055
  • Jul 1, 2019 3:15am Jul 1, 2019 3:15am
Quoting robots4me
{quote} _________ Couldn't we just make the indicator check every X candle or X time (ie every 30 M)? I know the exit touch would be late then, but we would survive that... PC
@pipcruiser -- what do you mean by "every X candle"? We make trading decisions based on the tick price -- now.[/quote]

_____

I really tried but I cant explain it better than I did already :o

I guess its kind of like in the old version, to keep the indicator for constantly "thinking" (using CPU) - it only "thinks" every ie. 30 min candle close...

Sorry for the childish explanation.

PC
Less is more...
  • #4,056
  • Jul 1, 2019 3:16am Jul 1, 2019 3:16am
Quoting skyway
{quote} In blackjack card counters bet low size when the count is low waiting for the cards to heat up to high count when they up their betsize. Likewise, when the signal is less efficient we ignore or trade with lower lotsize. Unlike blackjack, we don't have to manually do the count in our head. And since we have access to 28pairs 24/5 instead of this current deck that we track the count we have this luxury to do business when the edge is clearly with us. That's the huge advantage of forex over bj besides having to stay under the casino surveillance...
I did some work this morning on this and the result is posted on my thread. I adapted the concepts from this thread for the final product.
https://www.forexfactory.com/showthread.php?p=12358753#post12358753
  • #4,057
  • Jul 1, 2019 3:20am Jul 1, 2019 3:20am
Quoting bishopdotun
{quote} @R4M: I guess you missed this post. Just reminding you. There would be cases where exit is opposite band.
Quote
There would be cases where exit is opposite band.
@bishopdotun -- I recall seeing that. I didn't understand how that could be and, so, put that comment on the back burner.

I assumed the price always has to pass through the midline first. In your screen chart example it looks like the midline occurs shortly after you open a SELL. But the midline is still between the upper and lower bands.

Does this mean we need a rule that if the midline is too close to an upper or lower band then we ignore it and keep going until we reach the other side?

It seems we need a rule such that if the midline shifts while a trade is still open then the Exit becomes the opposite band -- something like that.

EDIT: However, on that same screen shot there is a later SELL signal where it makes sense to Exit at the midline, even though it has shifted.
  • #4,058
  • Jul 1, 2019 3:31am Jul 1, 2019 3:31am
Quoting pipcruiser
{quote} @pipcruiser -- what do you mean by "every X candle"? We make trading decisions based on the tick price -- now.
_____ I really tried but I cant explain it better than I did already :o I guess its kind of like in the old version, to keep the indicator for constantly "thinking" (using CPU) - it only "thinks" every ie. 30 min candle close... Sorry for the childish explanation. PC[/quote]

@pipcruiser -- please -- don't every apologize, otherwise I'll start ignoring you.

Actually, there are times (though not here) when I'll make use of MT4's built-in OnTimer function for exactly that purpose -- to give the CPU a break.

However in this case I think the strategy may need to process every tick. Or, perhaps, since we are using higher time frames it's good enough to check every 10 seconds -- something like that.

Perhaps @bishopdotun could provide some guidance.

EDIT: @pipcruiser -- now that I better understand your comment then I think you are correct. I wouldn't use M30 candles but I might switch to using OnTimer and check prices every 10 seconds. Compared to ticks arriving many times per second, this could work well. Good idea...
  • #4,059
  • Jul 1, 2019 3:37am Jul 1, 2019 3:37am
Quoting robots4me
{quote} _____ I really tried but I cant explain it better than I did already :o I guess its kind of like in the old version, to keep the indicator for constantly "thinking" (using CPU) - it only "thinks" every ie. 30 min candle close... Sorry for the childish explanation. PC
@pipcruiser -- please -- don't every apologize, otherwise I'll start ignoring you. Actually, there are times (though not here) when I'll make use of MT4's built-in OnTimer function for exactly that purpose -- to give the CPU a break. However in this case I think the strategy may need to process every tick. Or, perhaps, since we are using higher time frames it's good enough to check every 10 seconds -- something like that. Perhaps @bishopdotun could provide some guidance. EDIT: @pipcruiser -- now that I better understand your comment then I think you are correct. I wouldn't use M30 candles but I might switch to using OnTimer and check prices every 10 seconds. Compared to ticks arriving many times per second, this could work well. Good idea...[/quote]


________

Awesome! - Still think way less frequent would do the job....

The always apologizing is a "left-over" from being with my ex-wife LOL

PC
Less is more...
  • #4,060
  • Jul 1, 2019 3:42am Jul 1, 2019 3:42am
@R4M: Every 10s would be fine

You could make it an input
1 trader viewing now