Hello, Could you please tell me the formula to create an indicator to show standard deviation over the past 30 bars (including the current bar). I'm using 5 min bars and looking for volume that is greater then 1.95 standard deviation based on the 30 bars.

If I have the standard deviation plotted on my chart then I can create a scan condition that plot to look for SD greater then 1.95.

Thanks

The 30-period standard deviation of volume can be written as follows.

SQR(ABS(V ^ 2 + V1 ^ 2 + V2 ^ 2 + V3 ^ 2 + V4 ^ 2 + V5 ^ 2 + V6 ^ 2 + V7 ^ 2 + V8 ^ 2 + V9 ^ 2 + V10 ^ 2 + V11 ^ 2 + V12 ^ 2 + V13 ^ 2 + V14 ^ 2 + V15 ^ 2 + V16 ^ 2 + V17 ^ 2 + V18 ^ 2 + V19 ^ 2 + V20 ^ 2 + V21 ^ 2 + V22 ^ 2 + V23 ^ 2 + V24 ^ 2 + V25 ^ 2 + V26 ^ 2 + V27 ^ 2 + V28 ^ 2 + V29 ^ 2 - 30 * AVGV30 ^ 2) / 30)

Modelling Bollinger Bands (& Standard Deviation) in a TC PCF

Thanks Bruce, I created the PCF indicator and there was one formula error for parenthisis so it was obvious I just added a bracket at the end after 30.

But I'm wondering about scaling because I get values in the range of 100 to 500 K. I was expecting a normal distribution on a scale of maybe -10 to +10. Do you know what I mean. I could create a mathmatical standard deviation chart with sample data to help explain. The scale just seems way out there and I tried the scaling under edit with no luck.

Thanks

Bruce maybe an easier way to do this is plot a bollinger band on volume (set the Std Dev to 1.95 and period to 30) and then scan for volume crossing the top band.

I've done it and I can see when volume is crossing 1.95 std deviation but is there a formula that would return the actual std deviation  like 1.95, 2.3, 3.5 type of result? You see what I mean the actual standard dveiation some result like 500K

Thanks

Opps I meant "not some result like 500K"

You could divide the standard deviation by the average volume to make it comparable between symbols.

SQR(ABS(V ^ 2 + V1 ^ 2 + V2 ^ 2 + V3 ^ 2 + V4 ^ 2 + V5 ^ 2 + V6 ^ 2 + V7 ^ 2 + V8 ^ 2 + V9 ^ 2 + V10 ^ 2 + V11 ^ 2 + V12 ^ 2 + V13 ^ 2 + V14 ^ 2 + V15 ^ 2 + V16 ^ 2 + V17 ^ 2 + V18 ^ 2 + V19 ^ 2 + V20 ^ 2 + V21 ^ 2 + V22 ^ 2 + V23 ^ 2 + V24 ^ 2 + V25 ^ 2 + V26 ^ 2 + V27 ^ 2 + V28 ^ 2 + V29 ^ 2 - 30 * AVGV30 ^ 2) / 30) / AVGV30

Your idea of using Bollinger Bands is probably the best option for what I suspect you really want. I would probably use above instead of crossing up though.

If you want to plot how many standard deviations volume is above or below its moving average, you could use something like the following.

(V - AVGV30) / SQR(ABS(V ^ 2 + V1 ^ 2 + V2 ^ 2 + V3 ^ 2 + V4 ^ 2 + V5 ^ 2 + V6 ^ 2 + V7 ^ 2 + V8 ^ 2 + V9 ^ 2 + V10 ^ 2 + V11 ^ 2 + V12 ^ 2 + V13 ^ 2 + V14 ^ 2 + V15 ^ 2 + V16 ^ 2 + V17 ^ 2 + V18 ^ 2 + V19 ^ 2 + V20 ^ 2 + V21 ^ 2 + V22 ^ 2 + V23 ^ 2 + V24 ^ 2 + V25 ^ 2 + V26 ^ 2 + V27 ^ 2 + V28 ^ 2 + V29 ^ 2 - 30 * AVGV30 ^ 2) / 30)

Thanks Bruce, Sounds good and the  formula is returning some good results I can use.

HI Bruce, A bit a a simple question but how do I test if the current close is the max close for the Day. I wrote a Daily timeframe PCF (C=MaxC) but it doesn't work. It returns true all the time and just uses the current C value on both sides of the equation.

I just want to add a condition to check if the last 2 Min candle is reaching the Max C for the day.

I'm missing something simple.

Thanks

The maximum last/close price for the day would be the high, so C=H.

I don't think this will work becasue there could have been a high reached that is greater then the maximum close reached that day. So the maximum close would return FALSE. using C=H

I don't undertsand why the formula MAXC doesn't return the maximum close for the day. Shouldn't it. And then C=MAXC would return True if the close was the maximum reached so far that day.

No, MAXC should not return the highest close of the day. The format of MAXC in the Personal Criteria Formula Language is as follows.

MAXCx.z -> Maximum of Close/Last Price

Where x is the Period of the Indicator or Function.
Where z is the number of Periods Ago.
- z can also be a date. May 3, 2007 would be (single quotes required): '05/03/07'

PCF Formula Descriptions

So MAXC ends up just being C because not specifying a period means it defaults to 1.

You can check for price being at its high of the day currently with the formula given by StockGuy.

C = H

But you cannot check for price being at or above all of the previous closing prices during the trading day without knowing how many bars have happened so far during the trading day.

So for example, if you are ten bars into the trading day, you could use the following.

C = MAXC10

But also note that C doesn't become the close of the bar until the bar actually closes. Since we are not specifying a number of bars ago, this defaults to 0. This means it is looking at the current bar. Unless that bar is the last bar of the trading day and the market has closed, C is the current price and is subject to change until the bar closes. So prior to the end of the trading day, you would need to use C1 if you want to check for the actual closing price of the bar.

You would also need to increment the bars ago parameter in the MAXC section. So if you are interested in closing prices of intraday bars, you would need to use the following during the tenth bar of the trading day.

C1 = MAXC9.1

This would be checking for the close of the ninth bar to be at or above the closes of the previous 8 bars since only 9 bars will have actually finished forming.

Thanks Bruce and this is interesting that there is no way if your tracking 5 minutes bars to see if the last bar (C1) is the higest from all 5 min bars so far inthe day without specifying how many bars back. Too bad.

Thanks for clarifying.

I'm using daily chart and looking for volume that is greater then 1.95 standard deviation based on the previous 150 bars. Is that possible and could you tell me how to write that?

Thanks

The following Condition Formula is for volume being greater than 1.95 times the standard deviation of volume ending 1 bar ago.

V > 1.95 * SQR(ABS(V1 ^ 2 + V2 ^ 2 + V3 ^ 2 + V4 ^ 2 + V5 ^ 2 + V6 ^ 2 + V7 ^ 2 + V8 ^ 2 + V9 ^ 2 + V10 ^ 2 + V11 ^ 2 + V12 ^ 2 + V13 ^ 2 + V14 ^ 2 + V15 ^ 2 + V16 ^ 2 + V17 ^ 2 + V18 ^ 2 + V19 ^ 2 + V20 ^ 2 + V21 ^ 2 + V22 ^ 2 + V23 ^ 2 + V24 ^ 2 + V25 ^ 2 + V26 ^ 2 + V27 ^ 2 + V28 ^ 2 + V29 ^ 2 + V30 ^ 2 + V31 ^ 2 + V32 ^ 2 + V33 ^ 2 + V34 ^ 2 + V35 ^ 2 + V36 ^ 2 + V37 ^ 2 + V38 ^ 2 + V39 ^ 2 + V40 ^ 2 + V41 ^ 2 + V42 ^ 2 + V43 ^ 2 + V44 ^ 2 + V45 ^ 2 + V46 ^ 2 + V47 ^ 2 + V48 ^ 2 + V49 ^ 2 + V50 ^ 2 + V51 ^ 2 + V52 ^ 2 + V53 ^ 2 + V54 ^ 2 + V55 ^ 2 + V56 ^ 2 + V57 ^ 2 + V58 ^ 2 + V59 ^ 2 + V60 ^ 2 + V61 ^ 2 + V62 ^ 2 + V63 ^ 2 + V64 ^ 2 + V65 ^ 2 + V66 ^ 2 + V67 ^ 2 + V68 ^ 2 + V69 ^ 2 + V70 ^ 2 + V71 ^ 2 + V72 ^ 2 + V73 ^ 2 + V74 ^ 2 + V75 ^ 2 + V76 ^ 2 + V77 ^ 2 + V78 ^ 2 + V79 ^ 2 + V80 ^ 2 + V81 ^ 2 + V82 ^ 2 + V83 ^ 2 + V84 ^ 2 + V85 ^ 2 + V86 ^ 2 + V87 ^ 2 + V88 ^ 2 + V89 ^ 2 + V90 ^ 2 + V91 ^ 2 + V92 ^ 2 + V93 ^ 2 + V94 ^ 2 + V95 ^ 2 + V96 ^ 2 + V97 ^ 2 + V98 ^ 2 + V99 ^ 2 + V100 ^ 2 + V101 ^ 2 + V102 ^ 2 + V103 ^ 2 + V104 ^ 2 + V105 ^ 2 + V106 ^ 2 + V107 ^ 2 + V108 ^ 2 + V109 ^ 2 + V110 ^ 2 + V111 ^ 2 + V112 ^ 2 + V113 ^ 2 + V114 ^ 2 + V115 ^ 2 + V116 ^ 2 + V117 ^ 2 + V118 ^ 2 + V119 ^ 2 + V120 ^ 2 + V121 ^ 2 + V122 ^ 2 + V123 ^ 2 + V124 ^ 2 + V125 ^ 2 + V126 ^ 2 + V127 ^ 2 + V128 ^ 2 + V129 ^ 2 + V130 ^ 2 + V131 ^ 2 + V132 ^ 2 + V133 ^ 2 + V134 ^ 2 + V135 ^ 2 + V136 ^ 2 + V137 ^ 2 + V138 ^ 2 + V139 ^ 2 + V140 ^ 2 + V141 ^ 2 + V142 ^ 2 + V143 ^ 2 + V144 ^ 2 + V145 ^ 2 + V146 ^ 2 + V147 ^ 2 + V148 ^ 2 + V149 ^ 2 + V150 ^ 2 - 150 * AVGV150.1 ^ 2) / 150)

Can you tell me how to write the 30 period standard deviation of price rather than volume?

thanks

Just substitute C for all instances of V in the formula.

SQR(ABS(C ^ 2 + C1 ^ 2 + C2 ^ 2 + C3 ^ 2 + C4 ^ 2 + C5 ^ 2 + C6 ^ 2 + C7 ^ 2 + C8 ^ 2 + C9 ^ 2 + C10 ^ 2 + C11 ^ 2 + C12 ^ 2 + C13 ^ 2 + C14 ^ 2 + C15 ^ 2 + C16 ^ 2 + C17 ^ 2 + C18 ^ 2 + C19 ^ 2 + C20 ^ 2 + C21 ^ 2 + C22 ^ 2 + C23 ^ 2 + C24 ^ 2 + C25 ^ 2 + C26 ^ 2 + C27 ^ 2 + C28 ^ 2 + C29 ^ 2 - 30 * AVGC30 ^ 2) / 30)

Thanks Bruce...now if I wanted to incorporate the absolute value of open- close of the current candle with this 30 period standard deviation of price, how would I write that?

In what way? Do you want to add it?

ABS(O - C) + SQR(ABS(C ^ 2 + C1 ^ 2 + C2 ^ 2 + C3 ^ 2 + C4 ^ 2 + C5 ^ 2 + C6 ^ 2 + C7 ^ 2 + C8 ^ 2 + C9 ^ 2 + C10 ^ 2 + C11 ^ 2 + C12 ^ 2 + C13 ^ 2 + C14 ^ 2 + C15 ^ 2 + C16 ^ 2 + C17 ^ 2 + C18 ^ 2 + C19 ^ 2 + C20 ^ 2 + C21 ^ 2 + C22 ^ 2 + C23 ^ 2 + C24 ^ 2 + C25 ^ 2 + C26 ^ 2 + C27 ^ 2 + C28 ^ 2 + C29 ^ 2 - 30 * ACGC30 ^ 2) / 30)

Take the ratio?

ABS(O - C) / SQR(ABS(C ^ 2 + C1 ^ 2 + C2 ^ 2 + C3 ^ 2 + C4 ^ 2 + C5 ^ 2 + C6 ^ 2 + C7 ^ 2 + C8 ^ 2 + C9 ^ 2 + C10 ^ 2 + C11 ^ 2 + C12 ^ 2 + C13 ^ 2 + C14 ^ 2 + C15 ^ 2 + C16 ^ 2 + C17 ^ 2 + C18 ^ 2 + C19 ^ 2 + C20 ^ 2 + C21 ^ 2 + C22 ^ 2 + C23 ^ 2 + C24 ^ 2 + C25 ^ 2 + C26 ^ 2 + C27 ^ 2 + C28 ^ 2 + C29 ^ 2 - 30 * ACGC30 ^ 2) / 30)

Something else?

Perfect It was the ratio I was looking for..Thanks Bruce

Hi 

I am now looking to plot this:

ABS(O - C) / SQR(ABS(C ^ 2 + C1 ^ 2 + C2 ^ 2 + C3 ^ 2 + C4 ^ 2 + C5 ^ 2 + C6 ^ 2 + C7 ^ 2 + C8 ^ 2 + C9 ^ 2 + C10 ^ 2 + C11 ^ 2 + C12 ^ 2 + C13 ^ 2 + C14 ^ 2 + C15 ^ 2 + C16 ^ 2 + C17 ^ 2 + C18 ^ 2 + C19 ^ 2 + C20 ^ 2 + C21 ^ 2 + C22 ^ 2 + C23 ^ 2 + C24 ^ 2 + C25 ^ 2 + C26 ^ 2 + C27 ^ 2 + C28 ^ 2 + C29 ^ 2 - 30 * ACGC30 ^ 2) / 30)

on either the high or low of the candle whenever the value returns equal to or higher than 1.95...Is that possible?

Thanks

So maybe something like the following plotted as a Custom PCF Indicator set to the Dot Plot Style?

H * ABS(ABS(O - C) / SQR(ABS(C ^ 2 + C1 ^ 2 + C2 ^ 2 + C3 ^ 2 + C4 ^ 2 + C5 ^ 2 + C6 ^ 2 + C7 ^ 2 + C8 ^ 2 + C9 ^ 2 + C10 ^ 2 + C11 ^ 2 + C12 ^ 2 + C13 ^ 2 + C14 ^ 2 + C15 ^ 2 + C16 ^ 2 + C17 ^ 2 + C18 ^ 2 + C19 ^ 2 + C20 ^ 2 + C21 ^ 2 + C22 ^ 2 + C23 ^ 2 + C24 ^ 2 + C25 ^ 2 + C26 ^ 2 + C27 ^ 2 + C28 ^ 2 + C29 ^ 2 - 30 * ACGC30 ^ 2) / 30) > 1.95)

Or would is there some requirement as to when you choose the high vs the low instead of just plotting a dot on the high?

Thanks Bruce

Now if I wanted to look for this pcf indicator below but for the Previous 1 candle before it, how would I write this

h * ABS(ABS(O - C) / SQR(ABS(C ^ 2 + C1 ^ 2 + C2 ^ 2 + C3 ^ 2 + C4 ^ 2 + C5 ^ 2 + C6 ^ 2 + C7 ^ 2 + C8 ^ 2 + C9 ^ 2 + C10 ^ 2 + C11 ^ 2 + C12 ^ 2 + C13 ^ 2 + C14 ^ 2 + C15 ^ 2 + C16 ^ 2 + C17 ^ 2 + C18 ^ 2 + C19 ^ 2 + C20 ^ 2 + C21 ^ 2 + C22 ^ 2 + C23 ^ 2 + C24 ^ 2 + C25 ^ 2 + C26 ^ 2 + C27 ^ 2 + C28 ^ 2 + C29 ^ 2 - 30 * AvGC30 ^ 2) / 30) > 1.95 and  (V - AVGV30)/ SQR(ABS(V ^ 2 + V1 ^ 2 + V2 ^ 2 + V3 ^ 2 + V4 ^ 2 + V5 ^ 2 + V6 ^ 2 + V7 ^ 2 + V8 ^ 2 + V9 ^ 2 + V10 ^ 2 + V11 ^ 2 + V12 ^ 2 + V13 ^ 2 + V14 ^ 2 + V15 ^ 2 + V16 ^ 2 + V17 ^ 2 + V18 ^ 2 + V19 ^ 2 + V20 ^ 2 + V21 ^ 2 + V22 ^ 2 + V23 ^ 2 + V24 ^ 2 + V25 ^ 2 + V26 ^ 2 + V27 ^ 2 + V28 ^ 2 + V29 ^ 2 - 30 * AVGV30 ^ 2) / 30) > 1.95) 

Thanks

You would need to add 1 to all of the bars ago parameter is the formula. If you still want it to plot at the high of the current bar instead of plotting the high of the previous bar on the current bar you would need to change the H1 at the beginning of the following formula to an H instead.

H1 * ABS(ABS(O1 - C1) / SQR(C1 ^ 2 + C2 ^ 2 + C3 ^ 2 + C4 ^ 2 + C5 ^ 2 + C6 ^ 2 + C7 ^ 2 + C8 ^ 2 + C9 ^ 2 + C10 ^ 2 + C11 ^ 2 + C12 ^ 2 + C13 ^ 2 + C14 ^ 2 + C15 ^ 2 + C16 ^ 2 + C17 ^ 2 + C18 ^ 2 + C19 ^ 2 + C20 ^ 2 + C21 ^ 2 + C22 ^ 2 + C23 ^ 2 + C24 ^ 2 + C25 ^ 2 + C26 ^ 2 + C27 ^ 2 + C28 ^ 2 + C29 ^ 2 + C30 ^ 2 - 30 * AVGC30.1 ^ 2) / 30) > 1.95 AND (V1 - AVGV30.1)/ SQR(ABS(V1 ^ 2 + V2 ^ 2 + V3 ^ 2 + V4 ^ 2 + V5 ^ 2 + V6 ^ 2 + V7 ^ 2 + V8 ^ 2 + V9 ^ 2 + V10 ^ 2 + V11 ^ 2 + V12 ^ 2 + V13 ^ 2 + V14 ^ 2 + V15 ^ 2 + V16 ^ 2 + V17 ^ 2 + V18 ^ 2 + V19 ^ 2 + V20 ^ 2 + V21 ^ 2 + V22 ^ 2 + V23 ^ 2 + V24 ^ 2 + V25 ^ 2 + V26 ^ 2 + V27 ^ 2 + V28 ^ 2 + V29 ^ 2 + V30 ^ 2 - 30 * AVGV30.1 ^ 2) / 30) > 1.95)