Silicon Investor (SI) -- The First Internet Community

STOCKTALK

We've detected that you're using an ad content blocking browser plug-in or feature. Ads provide a critical source of revenue to the continued operation of Silicon Investor. We ask that you disable ad blocking while on Silicon Investor in the best interests of our community. If you are not using an ad blocker but are still receiving this message, make sure your browser's tracking protection is set to the 'standard' level.

Strategies & Market Trends : TA Science Projects & Experimental Indicators -- Ignore unavailable to you. Want to Upgrade?

To: Smooth Drive who wrote (164)	6/4/1998 10:07:00 PM
From: ftth	Read Replies (2) \| Respond to of 237

[Williams A/D] is defined as: ((C-O)/(H-L))*volume=daily A/D figure. (NOTE: Williams uses different A/D equations for different indicators; others besides Larry Williams utilize this definition of A/D (e.g. Tom DeMark, John Murphy) The most notable difference in this version of A/D is the use of opening prices. In candlestick terms, this indicator will contribute positive volume for white candle days (C>O), and negative volume for black candle days(C<O). The volume scale factor is the fraction of the daily range taken up by the candle body. If we open on the low and close on the high, AND the close was higher than yesterday's close, Williams' A/D contributes the same positive volume (ALL) as Granville's OBV. Same concept holds for opening on the high and closing on the low, and below yesterday's close. Like the other A/D measures, this is biased (i.e. larger than reality). However, the bias is lower that either Chaikin's or OBV. The use of all available data (H,L,C,vol AND OPEN) reduces the bias as follows: Whereas Chaikin's version assigns all the day's volume as an addition to the running sum on ANY day that closes on the high, the William A/D only assigns all the day's volume as an addition to the sum if the open is on the low and the close is on the high. Opens or closes in any other position contribute some fraction of the total available volume. Although the amount of volume assigned is still too high, it makes more intuitive sense to limit the high bias cases to days when we open on the low and close on the high (as opposed to ANY day that closes on the high, irrespective of where it opened). Opening on the low and closing on the high is the most positive scenario for prices, so if we can control our high bias case, this is where we would choose to apply it. One of the assumptions the Williams A/D makes is that the day's volume is uniformly distributed across the days range, and that the portion of the days volume that happened outside the candle body (i.e. in the shadows) consists of offsetting trades (buys and sells matched), and therefore doesn't contribute to the daily A/D figure. If you think about the mechanics of the equation, this is what it is implying. This makes intuitive sense as a way to make a first-order approximation to the volume distribution across the day's range, and how to divvy it up as either accumulation, distribution, or offset trades. We can make some refinements to this, but beyond a certain point the confidence in our assumptions goes way down (more on this later). Summary: Using the relation (C-O)/(H-L) as the baseline for an A/D indicator makes more intuitive sense than the other two A/D measures described in the prior posts. It is the only one that uses all available data (O-H-L-C-volume). The other two completely ignore perfectly valid data from an already-small sample of four prices per trading day. Neither of the other two measures accounts for the price movements from the open (and associated volume). Granville's uses only the closing price. As a user of these indicators, you have to decide whether you believe: 1. Only the closing price has significance, and the rest is noise; 2. Only the high, low and close have significance, and the open is noise; 3. The best results are achieved by using all available data; I choose the latter, knowing full well that using only four price points to represent an entire day's trading (possibly thousands of actual trades) is only a crude approximation. But it is a better approximation than the case of Chaikin's A/D or OBV. It also uses the four prices of greatest significance during a trading day. Any experienced user of candlesticks will confirm the value of using all four available prices for each day, compared to just 3 or just 1. This is the most convincing argument for me, because candlesticks were the single most significant tool I've added to my trading toolbox. The main thing candlesticks do is visually enhance the significance of the range from open to close, and the direction. I think Ray and Rainier would both confirm the value of using the four-price candlestick view compared to any other price tracking method. Some refinements to the Williams A/D can add even more of the information we can get from candlesticks. This may just give an A/D indicator with useful daily signals. dh