Ok, I posted some biotech tables over on Richard's threads but decided some of this data intensive stuff should come over here.
BTK closed at a 20 day high on Friday and I decided to test forward returns after such occurrences. I can shift it easily enough but for now am using a 50 day forward return to see try (word is try) to tease out if such an occurrence signals a bull move is underway and will continue. I would test a 1-10 day forward return if seeking to see if it signals overbought and a short term top.
The return profile is actually quite interesting and very unique. The main issue is that the answer to the above question highly (and I do mean highly) depends on the mood of the market. In bull markets, it appears being at the 20 day high is highly suggestive of sharply higher prices ahead. And in bear markets, sell sell sell. SERIOUSLY GET OUT!!!
I clearly manipulated the data to try to tease an overall trend out and could not. I am working with an abbreviated data set however and have removed cases with extreme new highs, and believe me there are some real doozies in there - thank you 2000 and 1998.
WITH the requirement that the market is hitting a new 20 day high, there are still two cases where the BTK is over 160% of the 20 day (yes 20 day) low. There are fully 9 cases above 50% and 15 above 40% (and if you count days not at peak 20 day levels, the cases increase). There are fully 43 cases above 30% above the 20 day low. Most of the periods with these extremes are either extreme bull markets or extreme bear markets - 2000 has top 10 occurrences, 1998 (thank you Russia) comes into play next, then the bull phase portion of 2009, then late 1999, add in some 2001 bear mkt rallies, some more new market runs in 2009, and some more 1998.
I was loving pivot tables as I became re-acquainted with them. I am hating them now. The base table allows you to filter out the extremes using >, <, etc. but the pivot table itself still refers to the full data sets and filtering there does not allow >, <, etc. but instead you must select individually each value to include. As I typed this, I just figured the work-around (create a new super rounded field that makes selecting values much easier). It also won't let me use a field as both a filter and row label (meaning it might be nice to exclude certain years from the analysis just to get an idea), but again typing this helped me (add the field again). My table is now pure data too now as sorting to try to tease out what to do gets destroyed by formulas. And I have lots of dummy fields everywhere as the functionality of the table lets me do some things but not others.
Anyway results are below.
Ugly formatting - apologies. I ran other variations and nothing really stands out here. Ran with ranges around 100 day MA and it did not help. Also ran vs. 50 day high - that seems to work a bit better but is highly variable by bull vs. bear. So tell me, are we in a bull or a bear in bios?
| | | | | | | When at 20 Day High | | | | | | | | When at 20 Day High | | | When at 20 Day High Excl Extremes | | | Ann. | Daily | Avg 50 Day | | | Avg 50 Day | | | | Avg 50 Day | | | Returns | Returns | Forward Ret | Occurr. | % Up | Forward Ret | | Occurr. | % Up | Forward Ret | | 1995 | 63.0% | 0.2% | 12.1% | 47 | 68.1% | 6.9% | | 34 | 73.5% | 9.0% | | 1996 | 7.9% | 0.0% | 3.8% | 32 | 68.1% | -2.1% | | 30 | 60.0% | -2.5% | | 1997 | 12.6% | 0.1% | 1.4% | 48 | 62.5% | -5.2% | | 41 | 29.3% | -5.5% | | 1998 | 14.0% | 0.1% | 4.2% | 28 | 27.1% | 5.3% | | 16 | 37.5% | -2.4% | | 1999 | 111.4% | 0.3% | 26.2% | 57 | 64.3% | 37.6% | | 49 | 79.6% | 33.3% | | 2000 | 62.0% | 0.3% | 4.5% | 36 | 82.5% | -2.6% | | 4 | 50.0% | -0.2% | | 2001 | -8.5% | 0.0% | -0.7% | 24 | 50.0% | -5.3% | | 6 | 66.7% | -2.9% | | 2002 | -41.7% | -0.2% | -7.4% | 20 | 45.8% | -8.1% | | 12 | 25.0% | -7.1% | | 2003 | 44.9% | 0.2% | 10.1% | 37 | 15.0% | 10.3% | | 31 | 64.5% | 11.6% | | 2004 | 11.0% | 0.1% | 0.0% | 37 | 62.2% | -1.2% | | 37 | 48.6% | -1.2% | | 2005 | 25.1% | 0.1% | 6.8% | 43 | 48.6% | 8.3% | | 43 | 97.7% | 8.3% | | 2006 | 10.8% | 0.0% | 1.5% | 36 | 97.7% | 2.8% | | 36 | 66.7% | 2.8% | | 2007 | 4.3% | 0.0% | -0.2% | 36 | 66.7% | -2.4% | | 36 | 30.6% | -2.4% | | 2008 | -17.7% | -0.1% | -2.0% | 23 | 30.6% | -7.9% | | 23 | 30.4% | -7.9% | | 2009 | 45.6% | 0.2% | 10.9% | 42 | 30.4% | 9.1% | | 33 | 51.5% | 10.0% | | 2010 | 37.7% | 0.1% | 5.3% | 49 | 61.9% | 6.5% | | 44 | 75.0% | 8.8% | | 2011 | -15.9% | -0.1% | 1.3% | 31 | 67.3% | -0.1% | | 31 | 64.5% | -0.1% | | 2012 | 39.0% | 0.1% | 5.5% | 55 | 64.5% | 2.8% | | 46 | 54.3% | 2.6% | | 2013 | 54.2% | 0.2% | 9.3% | 60 | 56.4% | 8.3% | | 60 | 91.7% | 8.3% | | 2014 | 47.5% | 0.2% | 7.6% | 62 | 91.7% | 7.2% | | 61 | 90.2% | 7.6% | | 2015 | 10.5% | 0.1% | -3.5% | 44 | 88.7% | -4.8% | | 44 | 40.9% | -4.8% | | 2016 | -14.0% | -0.2% | -0.6% | 9 | 40.9% | | | 9 | 0.0% | | | Grand Total | 18.9% | 0.1% | 4.6% | 856 | 62.5% | 4.8% | | 726 | 62.5% | 5.0% | |
