UnBelievable,
At this point it would be useful for me if Dick could reiterate the ways in which the data he is using and the analysis he does based on that data differs from mine.
Among the issues to clarify might be:
Does your analysis lead you to conclude that PMCS did not have negative money flow Friday?
Do you see more than 3 black 15 minute candles?
Do you agree that the volume during those periods does not appear to be above the average volume of a 15 minute segment.
As I believe I mentioned the IQCharts only provides money flow on a daily display so it is not possible for me to tell the money flow for each 15 minute segment.
Looking back through the thread for "Dick" and not finding him, I'm guessing this might have been meant for me. To answer the questions:
By the definition used by my charting service, money flow was not negative for PMCS on Friday. It generally trended higher throughout the day, with only a few small downturns.
I see 3 black 15 minute candles, apparently beginning at 10:00, 12:00, and 1:15 (the labeling on my charts is a bit suspect wrt to time, but this looks right- the time spacing should be correct.)
The volume during the first of those black candles was not unusual, fitting into the volume trend. The second and third were relatively light volume compared to the average candle volume for the day. The first two black candles had negative money flow, the third did not due to a long lower shadow, possibly the result of a bad print. Since the volume for that candle was small, it contributed almost nothing to the money flow anyway, so the long shadow does almost nothing to the day's money flow calculation. Two other candles had very small negative money flow. They were very short candles with small bodies and a slightly longer upper shadow than lower. This, and the third black candle are where the "Flow Factor" I talked about comes in. Because of this factor, a black candle with a long lower shadow yields positive money flow, and a white candle with a long upper shadow yields negative money flow. The sign and size of the factor depends on where the close is in relation to the high and low, with the mid point being zero, low = -1 and high = +1.
From other posts that define the WSJ version of money flow as being the difference between shares printing at or above ask minus shares printing at or below bid, I can certainly understand how a tick by tick calculation can deviate from the charts I see. I am not sure what to make of it, but one thing occurs to me that makes me wonder if the tick by tick version carries much weight. One can never know from looking at a time and sales record what the true sequence of transactions is, or the bid and ask price at the actual time of a transaction, and this is especially true for the block prints that make up a large % of the volume during the day. Large prints are notorious for being posted late, and the prices are often negotiated well before the transactions, so if the trend is up as it was for PMCS most of the day, these big prints are going to appear to be below the bid, even though they may have been buys.
Wishful thinking on my part? Not really. I spot checked some of the biggest prints in Friday's data and did not find any that printed at the bid. They were mostly lower, but if you looked back a few minutes, the prints there and the bid/ask are at or below the block print. I spent a little time looking at the T&S, and I sure can't find anything else that could account for the negative numbers obtained from any tick by tick method of determining money flow. I guess I 'll need to see something compelling to make me believe they are more credible than the methods I'm more familiar with that gave positive results for the day.
Dan |
