|Homework, Research, Statistics
The videos below were produced by Chad L. Payne of Linn Software as a series of homework lessons designed to address questions from the Investor/RT community. The videos are listed here in reverse chronological order. To view the videos in sequence, scroll down to the bottom of this article and work your way up to the top. After each video you will find links to chart definitions, setups that can be easily imported in Investor/RT to bring the chart into your Investor/RT setup. Note: This video on Exporting All Chart Data will be valuable to anyone doing their homework.
||Video 14: Timing of Events: New High and Low: In this video, we will look at the timing of certain events within the day. Specifically, we will look with daily bars at when the high and low of the day occurred, and color the bars based on which occurred 1st. We will also look at how to time other events such as where within the day the bar with highest volume occurred.
Additional statistics were computed on the timing of the high and low of the data after video was produced. The chart used for these statistics can also be imported below. Of the 2159 days tested, 990 of those days the high was set before the low (45.85%). This is as expected, since the ES moved up 600 points during that period and one would expect the high to be hit later in the day on up days. Looking at individual days, each day showed the high hitting last more than first, but Wednesday was somewhat of an outlier being more likely to hit the high first than the four other days.
HighLowTiming (1-min): [Definition]
HighLowTiming (Daily): [Definition]
HighLowTimingStats (Daily): [Definition]
|High and Low of Day Statistics
||% High First
||Days High First
||Video 13: Traveling The Initial Balance: In this video, we are looking at the 30 minute period following the Initial Balance (1st 60 minutes of session). During this 30 minute period, we are looking for price to move outside the initial balance and back into it. When this occurs, we are then looking to see if price traveled the full Initial Balance to the other side, before the session is over.
Of the 2055 days studied, our condition occurred 269 times on the IBH, and of those 269 times, 92 times we traveled back to the IBL by the end of the day (34.2%).
Of the 2055 days studied, our condition occurred 248 times on the IBL, and of those 248 times, 106 times we traveled back to the IBH by the end of the day (42.7%).
Overall, our condition occurred 517 times and of those 517 times we traveled across the IB by days end 198 times (38.3%).
Note: After importing charts below, user must use the buttons at the top of the chart to set Bars to 1, Start to 60, and Stop to 90, in order to receive results mentioned above. These buttons set V#20, V#18, and V#19 respectively.
SC Travel The IB (1-min): [Definition]
SC IBH To IBL (Daily): [Definition]
SC IBL To IBH (Daily): [Definition]
||Video 12: Timing Overnight High and Low Breakouts: This video focuses on the time of day at which price breaks above the overnight high or below the overnight low during the day session. We'll look at how often these breaks occur in the 1st 10 minutes of the day session, the 2nd ten minutes, and so on for each of the 41 10-minute periods of the day session. We'll do the same for 1-minute, 2-minute, 5-minute, and 30-minute intervals.
One of the more interesting discoveries in this video is the number of times the overnight high or low are broken just after 10am ET. This would likely be attributed to the release of reports at that time, but out of 2043 days, the ON high was broken between 10:30 and 10:31 45 times, compared to only 6 times during the 1-minute period preceding that. The complete data can be viewed in chart or text format using the links below.
1 Minute: [ Chart (png) ] [( Data (txt) ]
2 Minute: [ Chart (png) ] [( Data (txt) ]
5 Minute: [ Chart (png) ] [( Data (txt) ]
10 Minute: [ Chart (png) ] [( Data (txt) ]
30 Minute: [ Chart (png) ] [( Data (txt) ]
Time of Day Percentages Sheet: This shows the % of days on which the ON high or low have been broken as of each 10 minute period. It shows that, on 50.56% of the 2043 days studied (1033 days), the ON high was broken by 12:30pm ET. There were actually only 221 days in which the ON high wasn't broken until after 12:30, leaving 789 days in which the ON high was never breached during the day session.
(Import charts only after watching entire video. Some have very high computation times, and require the use of user variable V#19.)
||Video 11: Trend Days: This video looks at days where the high was made in the first 60 minutes and the low was made in the last 60 minutes, or visa versa. Of the 1900 days tested, 257 days (13.5%) made a high in the first 60 minutes and a low in the last 60 minutes. 372 days (19.6%) made a low in the first 60 minutes and a high in the last 60 minutes.
Trend Days Chart: [Definition]
||Video 10: Statistical Studies via Trading System Tools: This video describes how to flexibly produce statistical results using Trading System Tools. This method can be especially beneficial when several events are required to occur sequentially within a day or session. In this specific example, we look fro price to trade above the Initial Balance High (IBH), then retest then IBH. We then count the number of times these events occurred and priced closed above the IBH or below the IBH. The definition for the Trading System featured in the video can be downloaded below.
- 1895 days were tested.
- 1127 days price moved above IBH and then came back to test IBH
- Of those 1127 days, 617 closed above the IBH (54.7%) and 510 closed below IBH (45.3%)
Trading System: [Definition]
||Video 9: Change by Day of Week: This video compares the change in price by day of week and accumulates this data over a 600 week period.
- Over past 600 weeks, on average, Tuesday and Thursday have been very positive, Friday very negative, and Monday and Wednesday relatively flat.
Change by Day of Week Chart: [Definition]
|Change By Day Of Week
||Video 8: Opening Range Trades: This video discusses the how to setup chart to monitor and track the opening range trades.
- The video and results below reflect a 10-minute opening range. In order to test different opening range durations, you must open all signals ending in "_1MIN" as shown in the video, and adjust "POS > 10" to reflect the new duration, and also adjust the SESST_ORH and SESST_ORL tokens to reflect the new duration ("First 10 Minutes of Session"). You may actually plug in a user variable (i.e. V#5) into the 3 spots in all 6 indicators, and then simply change the V#5 value in one place (quotepage column).
- The Total Trades (and therefore the Percent Winners) are only considering those trades that hit the target or stop. Not including those that would have been entered but would have finished day without hitting either target or stop. The number in parenthesis is the number of trades that hit neither the stop or target.
- The original Opening Range Short Trades chart had an error in the calculation of the stopped out trades but has been corrected in the chart and definition below as well as the results.
- In these charts, to reach the target, we are looking for a move equivalent to the opening range. However, to reach the stop, we are looking for price to move from 1 tick above OR to 1 tick below. So technically, the stop is 2 ticks larger than target and would account to some degree to the hit targets outnumbering the stop outs.
Opening Range Trade Char (Long Trades): [Definition] [Image]
Opening Range Trade Char (Short Trades): [Definition] [Image]
Opening Range Trade Char (1-Minute Chart): [Definition] [Image]
*In the total trades, I am only considering those that hit target or stop. Not including those that would have been entered but would have finished day without hitting either target or stop.
||Video 7: Gap Statistics: How often does price gap up or down on the ES? And when it does, how often is that gap closed?
Gap Statistics Chart: [Definition]
Gap Statistics Plus Chart: [Definition] (This chart also shows how many days the gap was recovered by 50%. It also requires the user to set V#2 to the number of days to lookback for gap. This number will most commonly be 1. If V#2 is set to 2, it will look for a gap from 2 days ago instead of previous day).
Notes from Opening Statistics Video...
- Almost 1/3 of the time that price gaps 20 points in either direction, it closes that gap during the day session.
- Over 45% of the time price gaps 3 points or more, it never closes that gap and never returns to the previous close.
The study was repeated to include an upper limit on the gaps, resulting in statistics based on gaps within certain ranges.
Video 6: IOAMT Webinar: Video Recording of Webinar. The charts, topics, and statistics below were covered in the webinar. Chart and system definitions are provided for easy reproduction.
Day Session / IB Range Ratio: In this chart, we determined that the average ratio between the Day Session Range and the IB Range was 2.15. And on average, the Day Session Range was 10.38 points larger than the IB Range. These averages were computed on the ES over a 1,000 day period ending 5/12/2011.
Overnight to Day Session: In this chart, we determined that out of 1046 sessions, the days session broke out of both the high and low of the overnight session 279 times and stayed completely within the overnight range 44 times.
Volume Price Statistics: In this chart, we explored the Volume Price Statistics Indicator.
Volume Breakdown: In this chart, we explored the Volume Breakdown Indicator.
Return to Open: Here is the Trading System Definition for the Return to Open system discussed in the Webinar. Here is a screenshot of the optimization settings used to produce this spreadsheet.
Videos Mentioned in Webinar
Volume Breakdown Overview
Trading System Videos
Session Statistics Indicator
Signal Statistics Indicator
||Video 5: Opening Statistics: What is the maximum distance that price moves from the open of the day session each day? What is the average max move? What is the median max move? How often does price open in previous days range? When opening opening in range, how often does it touch both the previous days high and low? How often does it touch neither?
Opening Statistics Chart: [Definition]
Notes from Opening Statistics Video...
- These statistics are based on 1,200 day period of ES data ending 02/06/2011
- Average Move from Open: 14.71
- Median Move from Open: 11.50 (Mode: 7.00)
- Price Opened in Range 768 of 1200 days (64%)
- When opening in range, price touched both the previous high and low 77 of 768 times (10%)
- When opening in range, price touched neither previous high or low (stayed within previous days range) 124 of 768 times (16.14%)
- When opening in range, price touches either the previous high or the previous low (but not both), 73.86% of time.
||Video 4: VPOC Statistics: How do I export daily VPOC (Volume Point of Control) to excel along with open, high, low, close and volume for further analysis? How far on average was VPOC from previous day VPOC? How often did price open below/above previous days VPOC? What days had a VPOC very close to previous days VPOC? How often did price open in previous range and touch previous VPOC? How often did price open out of range and touch previous VPOC?
VPOC Statistics Chart: [Definition]
Spread Sheet of VPOC Stats: [htm file] [xlsx file]
VPOC - Volume Point of Control - Price at which more volume traded than any other price in the profile. In the context of this video, each profile represents one session. So the VPOC is the price at which the most volume traded each day.
Notes from VPOC Statistics Video...
- These statistics were computed over a 1230 day period ending 12/07/2010
- Average distance between consecutive VPOC = 11.68 points (Avg distance between lows = 11.68 also. Avg distance between highs = 10.06. Avg distance between closes = 12.35)
- Opened above previous VPOC: 621/1230 = 50.5% of time.
- Opened below previous VPOC: 595/1230 = 48.4% of time.
- Open on previous VPOC: 14/1230 = 1.1%
- How often did price open close to previous days VPOC?
- within 0.25: 46 (3.7%)
- within 0.50: 71 (5.8%)
- within 1 point: 108 (8.8%)
- within 2 points: 218 (17.7%)
- within 5 points: 468 (38.0%)
- How often did price open in previous range and touch previous VPOC? 557 (45.3%)
- How often did price open out of prev range and touch previous VPOC? 150 (12.2%)
- To change the chart to provide 24hr session numbers, simply change the chart to session 31 and the periodicity to 1425 minutes. All stats will then reflect 24hr session data, included exported data.
||Video 3: Average Range/Volume and Other Tendencies: What is the average range and volume of the initial balance, the overnight period, the day session, the 24hr session? How often does price open above/below previous day session range? What is the relationship of overnight direction to day session direction and visa versa?
I made a follow up video on 2/24/2012 which shows how to count the number of days where the IB was less than 4 and how to quickly compute the average daily range of those small IB days using a new Signal Statistics result: [Video] [Definition]
Average Range/Volume Chart: [Definition]
Open Out of Range Chart: [Definition]
Up/Down Relationship Chart: [Definition]
Spread Sheet of Sorted Ranges: [htm file] [xlsx file]
Notes from Average Range/Volume Video...
- These statistics are based on 1,000 day period of ES data ending 11/29/2010
- Avg Range: Day Session =20.25, 24hr Session=24.25, Overnight=13.75, IB=10.00
- Avg Volume: Day Session =1.74m, 24hr Session=1.97m, Overnight=236k, IB=404k
- Median Range: Day Session =16.75, 24hr Session=20.25, Overnight=11.25, IB=8.75 (Note: Median was not demonstrated in video - Uses Statistics Indicator).
- Median Volume: Day Session =1.68m, 24hr Session=1.92m, Overnight=217k, IB=397k
- Price Open Above Prev Day Sess Range: 247/1340=18.4% of time
- Price Open Below Prev Day Sess Range: 197/1340=14.7% of time
- Up/Down Relationship: Over a 1000 day period, the day session direction agreed with overnight session direction 29 more times than it disagreed. The relationship of day session direction to the following overnight session direction was insignificant (3 more times it disagreed over 1000 days).
|(Note: Min is questionable due to holidays and partial sessions.)
Average Range and Volume Per Bracket... (1000 Days ending 11/23/2010)
||Video 2: HOD and LOD Counts: How frequently does the HOD or LOD occur within each 30 minute bar/bracket? within each 5 minute bar? [Definition]
HOD - High price traded for the day / session.
LOD - Low price traded for the day / session.
Notes from HOD/LOD Counts Video...
- Results are based on 1,000 days of data ending 11/23/10, and only considering the day session HOD and LOD and day session bars on ES.
- HOD occurs in first 30 minutes of session 24.6% of the time. LOD 27% of the time.
- During the 2.5 hour mid-day period beginning 11:30pm ET, the HOD and LOD is rarely created (around 3% of the time for each 30-min period).
- HOD occurs in the first 5 minutes of the session 11.8% of the time. LOD 11.6% of the time.
- The likelihood of a HOD or LOD being created in first 5 mins of session is about the same as that of a 120-min period during middle of session. It might be interesting moving forward to make the same comparison w/ range and volume.
||Video 1: IB Breakout: How frequently does price break above the IB high, below the IB low, to both side, or stay inside the IB for the duration of the session? [Definition]
IB - Initial Balance - Range in which price traded for the first hour of the day session, typically 9:30am ET to 10:30am ET for the ES. IBH = IB high. IBL = IB Low.
Notes from IB Breakout Video...
- IB High Breakout Only = 36.6%, IB Low Breakout Only = 29.7%, Breakout to Both Sides = 31.6%, No Breakout = 2.1% (1,000 days ending 11/23/10)
|This is a video series titled "Homework, Research, Statistics". The videos above are listed in reverse chronological order. The videos answer question such as:
- How often did price extend both above and below the Initial Balance of the past 500 days? How often did it extend just above the high? How often just below the low? (Video 1)
- How often did price open above the previous day session range over the past 100 days? How often did it open below?
- What is the average day session range over the past 1000 days? What is the average IB range? What is the average overnight range?
- Frequency of HOD or LOD occurring within each 30 minute period. Within each 5-min period. (Video 2)
- Relationship of overnight direction (up/down) to following day session direction (up/down) and visa versa.
- If broke down market into 5-minute increments, breakdown of the times that a new high for the day has been made. Purpose would be to see if the chances of new highs being made during lunch period are very slim, while chances of new highs being made towards end of day are much greater. One year sample size.
- Same study as above, but new high would have to be made by 1.00 or more than previous swing high.
- # of lows of day that were formed by a double bottom and triple bottom over the past year (3-minute chart)?
- How many neutral, normal variation, normal days out of a given number of days.
- Average daily range over a number of days.
- Average IB range over a number of days
- How many days open with a gap (Out of range). If so, what was size of gap and then did the gap or ½ gap close
- If open Out of Range, then did it close back with previous day range.
- Number of points previous day VPOC and current day VPOC away from each other. Example, 1170, next day 1177, then seven points away.
- Identify double tops and/or double bottoms on any timeframe, paint/mark them as well and then e able to tell if it held or not.
- If Price opens X number of points away from previous day VPOC, how many points did it close away from previous day VPOC. Example, if open >7 points away, then did it close closer than 7 points or further than 7.
- How often did price open below/above previous days VPOC?
- How often did we have a gap open with a continued drive in direction of gap of say first 2 hours
- How often did we have a gap that got closed same day
- How often did we have a gap within previous days range
- How often is the first hour's range contained with the second hours range- (could be important for ex FDAX, FESX where cash opens 1 hour after initial open, to establish which IB is the most appropriate to use)
- What percentage of todays volume is contained within first 2.5 hours?
How often do we have trend days using following definition : open close to low and close is near the high
- How often do we have days where IB is violated on both sides with a close within IB, with a close on one of the extremes?
- What is the rotational factor of respective product?
- What days opened at or very close to previous days range extreme?
- What days opened in the middle of yesterdays range
- What days opened below/above previous days value area?
- What days had a VPOC very close to previous days VPOC
- Being able to find statistics on Opening Types and how they relate to Day Types would be helpful.
Open Auction In Range - Opens Inside Yest Range
Open Auction Out Of Range - Above Yest Range
Open Drive - Drives and doesn't retest open by more than a tick. Also it respects the swings and doesn't cross previous swing hi or lo.
Not easy to quantitate
Open Test Drive - Moves to test a prior area and
Open Rejection Reverse -Moves in the direction it was headed and reverses.
Normal - Range is bound by IB
Normal Variation - Breaks out above either IB hi or Lo not both
Neutral - Breaks both IB hi and Lo but settles back inside IB
Neutral Extreme - Breaks both IB hi and Lo but settles at one extreme Non Trend - Holiday (not relevant)
So there's endless cross reference studies that could be done. I'm most interested in finding out about OAIR and OAOR opening types and typically what kind of day they turn out to be. Also, IB size, extension above/below IB, where did it close in relation to open or hi/ lo.
- How often price stop and reverse when it approach or hit:
CHVN vs Yesterday HVN, CLVN vs Yesterday LVN, NVPOC, Yesterday (RTH) close vs Settlement
- I am learning excel. First question is how do I build a database. I was going to go thru the charts by hand for many of these data points.
- What is the effect of the o/n range on the days range? average o/n range, above and below average, extreme narrow and extreme wide o/n range.
- What is the effect of the above on the open type itself, tradable probability of open type from the above?
- Probabilities of different opening types based on open, gap up, down, above value, below value.
- Probability of day type following open type.
- Be able to extract volume for an opening swing/range to compare it to other days opening swing/range. If using your own opening swing, then would be some way to be able to pick that area and grab that volume, or be able to pick the time it took for that swing and grab that volume.
- What is most common daily range over past 250 days?
- How often did price extend just above the IB high? [to expand on this, what is the most common distance price extends past the IBH over that period of time]
- How often did price extend just below the IB low? [to expand on this, what is the most common distance price extends past the IBL over that period of time]
- For every day (historically) the ability to list Open price and then what NVPOCs and NCloses are touched and at what time. So the text file would deliver both the price touched and time.
- Another thing that I do is break down Opens and Closes into zones numerically. For Open, if we are above HOD=1, HOD>X>Value Area HI=2, if X is in upper 23% of value area =3, X is in middle 23% of value=4, bottom 23% =5, LOD<X<Value area LO=6, X<LOD=7............ Then I repeat this for the CLOSE based on End of Day Profile. Obviously if I am basing it on EOD profile, I can't have a value of 1 or 7 because I can't finish above or below value.
- On days that the price breaks the IB High/Low, how many ticks did it go past the break before it came back and hit the IB high/low? Is it possible to write a study like with that additional information. This could then be used to look at breaks of any defined level and how far the market went after the break.
- When ES gaps > (x) pts, it closed up/down (x) times the following session, and created a higher high or lower low (x) times.
When ES creates an inside day, how frequently does the following day gap up or down? How often is that gap filled?
How often does an expanded range (outside first standard deviation) lead to an inside day?
Others are all related to trade specific setups, ie when close is also VPOC, how often does it provide support/resistance? Not sure if raw data from IRT can provide this? Perhaps Fractals?
- Number of days out of a set number of days ES opened in previous days range and touched previous days VPOC.
Number of days our of a set number of days ES opened out of previous days range and touched previous days VPOC.
What % of the time is VAH broken if we open above VAH? and if it is broken what % of the time is POC touched? and if it is broken what % of the time is VAL touched? The opposite for VAL (http://www.charthub.com/images/2011/02/17/Homework_VA)
how often gap (1 pt or more) filled when opening in previous day range (1396 days studied)
- when opening out of range
- when opening above an up days range (up day defined as close greater than open) - 119/185
- when opening below an up days range - 28/47
- when opening above a down days range - 28/38
- when opening below a down days range - 90/120