Luger 

For an uptrend day: open within X% of day low close within X% of day high low occurs during first Y hours of day high occurs during last Z hours of day Luger

This gets my vote. Simple and will cover 90% of actual trend days.

Fat Tails 

This is my definition of an uptrending day: - The market opens near or at the low of the day. - The market closes near or at the high of the day. - The daily range is above average. - Throughout the day there is absence of sellers. The third condition needs to be explained. Profit taking may cause prices to retrace after a strong upmove, but if there are no sellers the move down cannot be sustained, but will be seen as an opportunity to buy. Buying usually takes place at the VWAP or above the VWAP. Therefore I would translate the third condition into -> prices not being able to exceed the lower noise level after open -> prices staying above the VWAP all day long after the initial balance On a strong uptrending day, prices will mostly remain between the first and second upper standard deviation bands of the VWAP. Vice-versa for downtrending days. Example: March 13 for ES 06-12 Example: March 20 for CL 05-12

same could be said for the 20sma, PA retraces to it but never breaks it. but I perfer Vwap now

Fat Tails 

True Range must be larger than ATR(3) of preceding day: TR[0] > ATR(3)[1]

What if we have three trending days in a row? They don't need to each be "bigger" trends than the prior day to still be considered trends when compared to a range bound day.

Perhaps we should be defining a range bound day.

Mike

save you the trouble...

Big Mike 

What if we have three trending days in a row? They don't need to each be "bigger" trends than the prior day to still be considered trends when compared to a range bound day. Perhaps we should be defining a range bound day. Mike

trend days do not have to be preceded by a day that has experienced a range contraction to be considered a trend day

there just so happens to be a 62% chance that it will be preceded by one, however

for statistical measurement you just have to decide on certain parameters to qualify your data and go from there

in a low volatility environment <10 pts may suffice as trend day range, but in a high volatility environment you might want to consider 15 points or 20 points range as a minimum, etc.

in it's broadest sense The market will open near its low price for the day session and build its way higher through the day, closing near its high price. The market will tend to stay above its VWAP for most of the day; for an uptrend day - and for a downtrend the converse.

if you want to get more specific , you can add other benchmarks for qualification, i.e., the $TICK, cumulative delta, pivots, etc.

tigertrader 

save you the trouble...

Do you know the exact method used to calculate those results?

That is what I am trying to accomplish --- the source of the methodology used by "the guys" who come out with these type of session statistics. I want to generate my own so I can fully understand and verify the data.

Mike

https://www.verticalsolutions.com/download/Little%20Book%20of%20Trend%20Days.pdf



Vertical Solutions

.

Why does so many include prior day data in the definition of a daily trend? What does the prior day's anything have to do with today's trend?

GridKing 

https://www.verticalsolutions.com/download/Little%20Book%20of%20Trend%20Days.pdf Vertical Solutions

Did not know this paper. Astonished that he also chose 20% as I have suggested. However, he only compares the range of trend day with the range of the prior day, whereas I have taken the average of the prior 3 days. But very close to what I was thinking...