AccDist /w Stochastic boundaries + CCI Extremes Strategy

RickK

Active member
Warning: This is a repainting indicator/strategy. Proceed with caution.

So I keep coming back to the AccDist (with Stochastic) and the AccDist (with Linear Regression) and I'm starting to really like it.
The original idea came from a post thread that begins here: https://usethinkscript.com/threads/a-wonderful-simple-top-and-bottom-intraday-signal.1981/ and
I posted something awhile back here: https://usethinkscript.com/threads/a-wonderful-simple-top-and-bottom-intraday-signal.1981/post-39357

So I've patched together a little strategy that finds excellent entries and definitely successful exits... but not optimal exits. I'm hoping that others might jump in and start a dialog about how to optimize exits. What I did was combine the AccDist_Stochastic indicator with a CCI_Extremes indicator (which basically just signals when CCI has hit +200 or -200....which are often reversal or retracement points).

Note: I've come to the conclusion that with either of the AccDist indicators, displaying them on charts with multiple days can be deceiving. Meaning that the indicator may display wonderful tops and bottoms on, let's say, a 5min/10 day chart, but when those days were happening in real time, the signals may not have been there. The indicator is fickle that way. So, my conclusion is this: use this only on a "today" chart, meaning 3min:Today, 5min:Today, 10min:Today, 15min:Today and so on.

A lot of the entries happen after market closes/reopens, overnight and quite often 1-2 hrs before market open.

I didn't note original authors in the script, so let me say that all credit goes to @BillMurray , @Playstation @horserider , @BenTen , @mashume , @we29125 and others. I was just monkeying around tossing snippets into someone elses code.

I have included a chart layout that has the strategy (2 slightly differerent versions) in side by side charts like the image below. That chart is located here: http://tos.mx/VRj7S2b

Thanks to any and all that jump in.

uXDkmoN.png
 
Last edited by a moderator:

andre.muhammad

Member
VIP
Great work here! I saw that post about two weeks ago and I started it using it. I must say it has helped greatly with entries too! I like what you have here and will test it out on Monday!

Your Acc/Distr is bounded, yes?
 

RickK

Active member
Great work here! I saw that post about two weeks ago and I started it using it. I must say it has helped greatly with entries too! I like what you have here and will test it out on Monday!

Your Acc/Distr is bounded, yes?

Yes, it seems to be bound by the Stochastic range 0 - 100, even though the AccDist range is 0 - 1. I tried eliminating the whole Stochastic mechanism, but it wouldn't work. I don't display the stochastic part in the lower indicator... I'm just interested in having the AccDist tied to its range. That said, the AccDist OB is 1.0 and OS is 0.0.
 

Bobbydigital83

Member
VIP
If if the the chart was set to "today," but lets say TSLA kept have multiple highs for the day, it would still repaint short entries and vice versa for long entries? Will the lower indicators repaint their "extreme high" and "extreme lows" as well?
 
Last edited:

RickK

Active member
If if the the chart was set to "today," but lets say TSLA kept have multiple highs for the day, it would still repaint short entries and vice versa for long entries, correct? Will the lower indicators repaint their "extreme high" and "extreme lows" as well?

Ha! That's the question. Sometimes it repaints and sometimes it doesn't. Actually, the way the code was written before was that when AccDist rolled UP over overbought, the signal would display. I changed that to: when AccDist rolled DOWN over overbought. That way there is at least a retreat in progress. Often I find that it will give a signal (let's say on a 10 min chart) and as soon as it hits the next candle, if there is not another arrow, it is a good time to take a scalp off the 1 or 3 min chart. Check out the image below and look at my post above for info about repainting signals and having successive signals.

I actually use the AccDist_Stoch and the AccDist_LinearRegression indicators together. Neither is necessarily better than the other but when they hit at the same time, its usually pretty good. .... or at least, conditions are right for a turnaround.... in the next minute, or ten or thirty...but it's coming.

Iq7aUTG.png
 

Piper2808t

Member
2019 Donor
VIP
thanks for sharing . I also used and have followed the author deafdaytrader on twitter until his death, r.i.p. with the strategy you created i am assuming this will not repaint? so in real time with say 3 min today the signals will be accurate and the arrows will not disappear with higher highs?
 

RickK

Active member
If if the the chart was set to "today," but lets say TSLA kept have multiple highs for the day, it would still repaint short entries and vice versa for long entries? Will the lower indicators repaint their "extreme high" and "extreme lows" as well?

I wanted to go back and see if what I was saying was actually so. I must note that in the image above, I highly doubt that if the chart was set to 10min:TODAY you would see that string of red down arrows across the top of the indicator. Here's a case in point. I just took a snapshot of SPY.... left side 5min:3DAYS and the right side is 5min:TODAY. Because the indicator is drawing from 3 days of data, AccDist is not considered as extreme a high as it is when only taking data from the current (globex and RTH session).

eMR4lXd.png
 

RickK

Active member
@Piper2808t Best I can say is read my post from the link at the top of the page and..... well, test it out. With multiple time frames AND comparing results of the SAME time frames but different sample periods (i.e. 5min:3DAY vs. 5min:TODAY)
 

RickK

Active member
@RickKennedy Have you noticed any repainting for the entry signal on the "TODAY" time frames?

I don't believe so, but I'll have to check. Note that there is no true repainting per se. If an alert is triggered on candle 1 (arrow) and an alert is triggered on candle 2 as well, then there will be 2 alerts shown....not a removal of the alert on candle 1 to be replaced by the one on candle 2. That's what repainting is as I understand it.

That said, I recall that there were lots of successive candles when the code called for an alert when the AccDist simply 'hit' 100 (overbought) but now that the alert is triggered when AccDist 'rolls back under' overbought, what will happen (at least on a non TODAY chart) is the absence of an alert (arrow). When this absence happens on a 10min or 5 min candle, it is often a good chance to take a quick scalp from the 1min chart.... even if price then rises again and throws another alert on the next 10 min candle. It's the absence that I look for to take the trade. The arrows simply tell me that it's coming. (It's definitely best to use this with other indicators to see if there are divergences, S/R areas, etc.) However, I haven't confirmed repainting on TODAY yet, but I suspect the same sort of thing. Keep in mind that it works quite similarly to any other oscillator and that "overbought" is a bad name that has been given to "strong strength in buying". And this strong strength can and does often continue for awhile. It's when the indicator rolls back down out of "overbought" that things 'might' be changing.

The key point to remember is that the TODAY chart is only looking at extremes from data from the beginning of trading on that specific day. I don't know why but for some reason, it sometimes seems to give different (more reliable?) signals than if you were looking at a non-TODAY chart. It's fun to set your chart to 10min/30 days and see all of the alerts that resulted in awesome moves ('after' 1 or successive candles), but as I said, I've concluded (or am concluding) that when those days were happening in real time, the signals may not have been there.
 

samoya

Active member
VIP
I added the Cumulative RSI MTF 1H to the 5M chart.
I think it could be considered to use it for the exit.
Long trade: exit after the CumRSI 1H moves the 1. step lower
Short trade: exit when the CumRSI 1H moves the 1. step higher
What do you think?

hxX7uvX.png
[/IMG]


Link for the Cumulative RSI
https://tos.mx/VoBQbjX
 

jnitro

New member
VIP
I took out CCI and replaced the following...looks promising on a 15min timeframe:

def Short1 = AccDist crosses below ad_over_bought;
def ExitShort1 = AccDist crosses above ad_over_sold;

IvTXdxO.png
 

RickK

Active member
@RickKennedy Have you noticed any repainting for the entry signal on the "TODAY" time frames?

New answer: Just now an hour after the Sunday open, I noticed that the indicator does literally repaint. That doesn't make it a dealbreaker tho, IMO. Just need to rely on something to confirm the reversal or trend change.

And the strategy actually goes back and 'uncloses' exits, which I guess is repainting too. So the strat performance $'s can't necessarily be relied on.
 
Last edited:

RickK

Active member
New answer: Just now an hour after the Sunday open, I noticed that the indicator does literally repaint. That doesn't make it a dealbreaker tho, IMO. Just need to rely on something to confirm the reversal or trend change.

And the strategy actually goes back and 'uncloses' exits, which I guess is repainting too. So the strat performance $'s can't necessarily be relied on.

Try this. Put two the two different AccDist indicators on the same chart. When both throw off a signal at the same time, it is probably a good one.
Here's the second one: http://tos.mx/5fZ2etK
 

FOTM_8888

Member
VIP
I took out CCI and replaced the following...looks promising on a 15min timeframe:

def Short1 = AccDist crosses below ad_over_bought;
def ExitShort1 = AccDist crosses above ad_over_sold;

IvTXdxO.png
hi. now the strategy have more profitable.. look very good with out eh CCI, can you share the complete strategy here with us please, so we can backtesting. thank you in advance.
 

andre.muhammad

Member
VIP
RickKennedy, I noticed the repaint on SPY today as it continued to repaint from the open to 12:00 PM EST today. I noticed that the indicator that used the linear regression never gave a sell signal... I just thought I would share that with you. Thanks!
 

Similar threads

Top