|
PureBytes Links
Trading Reference Links
|
More discussion...
| -----Original Message-----
| From: Neil Harrington [mailto:njh@xxxxxxxxx]
| Sent: Wednesday, June 03, 1998 3:27 PM
| To: bruceb@xxxxxxxxxxxxx
| Subject: RE: Some S&P Ideas & Questions
|
|
| Bruce,
|
| Thanks for your thoughts. I haven't done any stringent, formal
| analysis. It comes more from common sense and thinking things through.
|
| Some observations:
|
| 1. The growing trend in the world seems to be towards global
| electronic trading. Some other traders I know have started
| researching correlations between foreign "S&P" composites and
| the S&P, and their initial research shows a relationship. These
| points leads me to think that the night session will become
| increasingly significant.
|
| 2. All documentation of trades, whether they happen in the night
| or day session, show a history of price action and volume, which
| is the raw material of technical analysis. If you are missing
| any of that history, just like if you have bad data, your
| analysis is invalid, or at least based on incomplete or
| incorrect data. You wouldn't trust your analysis or chart
| patterns if you had a two hour gap in your day session.
|
| 3. I am not a big believer analyzing gap openings, since they
| never just happen all of a sudden (to any significant degree).
| They happen as a result of consist price action during the night session.
|
| 4. Because the night session is sparsely traded, if you are
| looking at x minute bars, the patterns are indiscernible, and
| indicators seem broken. But if you look at y tick bars, the
| patterns and indicators are as discernable and effective as
| during the day session. It just evolves over a longer period of
| time. If you like to look at 30 minute bars, but want to try
| ticks, just look at the average of how many ticks happen every
| thirty minutes during the day, and use that value for the number
| of ticks.
|
| 5. I cannot refute or comment on your research of the gaps and
| flag patterns that you have analyzed with and without the night
| session, but it seems that for your analysis to be complete, you
| should also do analysis on intra-day data with a two-hour gap of
| missing data. I don't mean to be facetious with that comment. As
| I have thought about this issue more, it has just become more
| apparent to me that if you don't look at the whole trading
| history, you have gaps in your data, you have gaps in price, and
| you have broken indicators until enough time passes (based on
| how far back your indicators look). I would agree that a two
| hour gap of data during the day session is much more detrimental
| that a two hour gap in the night session, but the general
| principle still holds, IMHO.
|
| I have many requests to post this thread to the list. Do you
| mind if I post this to the list?
|
| Thanks for making me think through my ideas. Please reply with
| your thoughts.
|
| Neil
|
| | -----Original Message-----
| | From: bruceb@xxxxxxxxxxxxx [mailto:bruceb@xxxxxxxxxxxxx]
| | Sent: Wednesday, June 03, 1998 2:20 PM
| | To: Neil Harrington
| | Subject: RE: Some S&P Ideas & Questions
| |
| |
| |
| |
| | > -----Original Message-----
| | > From: Neil Harrington [mailto:njh@xxxxxxxxx]
| | > Sent: Monday, June 01, 1998 1:25 PM
| | > To: ATI List; omega-list@xxxxxxxxxx
| | > Subject: Some S&P Ideas & Questions
| | >
| | >
| |
| | Neil, in regards to you statements below, I was wondering if by
| | any chance
| | you had done any "hard" analysis of the effect of the overnight
| | S&P session
| | on the day session.
| |
| | I ask this because your thoughts about the difference between
| a 200 point
| | move and a 500 point move that dropped to 200 before the open
| | appear to make
| | perfect sense on the surface, but it has been my observation
| | that the only
| | relevance on the day session from the night is the actual gap on
| | the open.
| |
| | By this I mean it just doesn't seem to make any difference what
| | the actual
| | "action" is in the overnight session, a 200 point gap open is
| a 200 point
| | gap open. I've reached this conclusion purely by visual
| | observation, that
| | is why I'm interested in any actual tests that any trader has
| | actually done.
| |
| | I have, however, followed one particular piece of evidence.
| I'm not the
| | biggest believer in classic chart patterns, but I have watched
| | the S&P for
| | simple bull and bear flag patterns for some time now in time
| | frames from 30
| | min on up. As I'm sure you're aware, the theory states once the price
| | breaks out of the bull "flag," it should make a move upwards at
| | least equal
| | to the size of the original "flagpole."
| |
| | Now the merit of this signal can certainly be debated, but the
| | interesting
| | thing to me is that the action in the overnight market seems
| to make no
| | difference on the outcome of the trade. For instance,
| suppose you have a
| | bull flag with a flagpole of +500 points from the base, a flag
| | top of +480,
| | and a flag bottom of +380 (prices have basically oscillated
| | between +480 and
| | +380 since the explosive move), and the market closes on the day
| | session at
| | +420.
| |
| | During the night session, however, the price breaks above +480.
| | The theory
| | says the target should be +1000, but it only goes up to +650 and
| | then drifts
| | back down to +450. If the night session really has an
| impact, this would
| | look like a great opportunity buy at the open the next day, because
| | (theoretically) the market has "tipped its hand," the
| breakout above +480
| | has already occurred, the theory's protective stop of +380
| has not been
| | violated to the downside, and you're actually being given a
| | chance to get in
| | 30 points below the initial breakout price.
| |
| | Unfortunately, my visual observations tell me this just isn't
| so. These
| | trades win no more often than ones where there was no breakout in the
| | evening session. This also appears to be the case on the flip
| | side as well.
| | That is, suppose during the night session (using the same
| | example) the price
| | did soar to +1000 and hit the target, and then collapsed back to
| | the trading
| | range before the open of the day session.
| |
| | You would think that any bullish breakout during the next day
| | session would
| | be less likely to result in a successful move, because the target was
| | already hit during the night session, but this just doesn't
| | appear to be the
| | case. In other words, bull flags appear to work just as
| often when the
| | target price was hit during the evening session as when it
| | wasn't. The same
| | seems to hold true for bear flags, so the historically upward
| bias of the
| | S&P market isn't a factor.
| |
| | The only exception to this seems to be when there is a VERY
| strong move
| | between 8-9:15 EST, but that action is primarily related to
| US government
| | and corporate morning reports. Any "momentum" this might
| cause is usually
| | fully known, and therefore doesn't really present an opportunity.
| |
| | The whole idea of using the evening session as an indicator
| for the day
| | session is very intriguing. Unfortunately, I just haven't seen
| | any way to
| | do it yet! Please let me know if you have any evidence to
| the contrary.
| |
| | Thanks,
| |
| | Bruce
| |
| | > 1. I use a BMI satellite feed and just trade the day session, but
| | > I use the
| | > SP8M (day and night session combined) symbol, for example,
| | rather than the
| | > DSP8M (day session only) symbol. I do this because it makes
| | the charts and
| | > indicators more continuous at the beginning of the day
| | session, and to me,
| | > the data more relevant to whether things are really more bullish
| | > or bearish
| | > at the beginning. For example, if you were just looking at the
| | day session
| | > only, and it gapped open 200 points higher, you would think
| things are
| | > definitely bullish from looking at the charts. But if were
| | displaying the
| | > night and day session together, you might have a different
| | view. Maybe in
| | > the night session an hour before the day session opens, the S&P
| | > was actually
| | > 500 points higher and the last hour before the close of the
| | night session,
| | > the market slides from +500 to +200 points higher than the close
| | > of the day
| | > before. Now your chart pattern is more bearish than bullish. To
| | > me, being a
| | > short term day trader, this last view is more reflective of
| the current
| | > sentiment when the market opens for the day session.
| |
|