Here is what I do with to spot sector trends and rotation. 1 TC2000 has 239 sectors which I track on daily basis. 2 Use a weighted relative strength formula to identify fast moving sector ( (2 * C * 100 / C5) + (1.5 * C * 100 / C25) + ( C * 100 / C40) ) /4.5 This quickly spots a new sector move. I put this in a custom ranking which with some color coding gives you a visual view so sectors with recent ranks are green and then it becomes yellow and red so it is easy to visualize the same. 3 I Also track top 100 and bottom hundred stocks using the same. And that list I use to find out sectors, this is basically because in some sectors there are hundreds of companies and only few start the first moves. 4 Track sectors which are making all time high as these sectors show strong trend after that.
Why not take it a step further? Make a chart of all those sectors, 3 months lets say. Put them on the same chart showing % gain/loss. Buy the top performer short the worst. Close the positions in a year, let's say. You will have had a long position in the top performer for the year, and a short position in the worst performer. By starting with a 3 month chart, you won't be rotating daily because by then some clear leading and lagging sectors will reveal themselves and will only, if at all, be crossed over a few times over the next year. Only way you can lose is if the top performing sector is a loser (or a very small winner) or nothing does particularly well and you end up rotating positions to stay with the top performer weekly or even more often in which case you will slowly fall away from the performance of the top sector due to coms/slippage/etc. Think about it
Thanks a ton, Grob. I really appreciate your time and efforts you put into your explanations! I see quite some logic in your reasoning that volume increases silently at the beginning, as people with "higher interests" prepare themselves for the next sector shift. There's much value to think about in your writings, and I will go thru the ideas and thoughts. Thinking about the data, I wonder if it is better to construct a equal dollar weighted or a cap-weighted sector proxy for my analysis. Equal-dollar weighted means a fair look at every stock in a sector. But are "people with higher interests" fair? They would prefer high cap stocks with high free float, as not to make too much noise, won't they? Again, thanks to you. This is going to be quite interesting! agrau
It depends on whether the sector breaks to all time high after a multi month base or basically it breaks after a furious rally from a significant correction. Like last week the lodging sector broke to all time high now there is a base , a cup and handle pattern and good volume , so I built positions in it and added to CHH which I was holding for over a year. As against that aluminum sector made an all time high after a multi month rally and just corrected after that. Also individual issues can have significant corrections and you don't know whether that correction is going to resume the trend(e.g. I had a position in USNA for many months and it gave me over 300% return and then had severe correction but the sector continued to consolidate. The sector rotation cycle also has some peculiarities some sectors just keep going without correction while some have three month rally and then they just have sideways move( e.g. the book publishing sector recently). So individual stock selection becomes a key and you have to keep ditching non movers and replace with new stocks with velocity. Sometimes I end up holding stocks for months sometime a sector makes a blowout move in few weeks like Security sector. So you have to play it accordingly. What works on long side for sector does not work on short side. Shorting the sectors at bottom is not the best of the strategy.
Regarding shorting, sure, you can throw that out. Do you realize what I'm saying though? Regardless of the reason for why the sector we are long is performing better than the others, we hold it until a different sector rises above it in % gain from the point we did the picking (and then minus an additional 3 months because of the original look-back period). If you keep rotating and switching to the sector that has had the best % return to date, you will have had it's performance as well. As you know sometimes different sectors rotate in and do extremely well like you gave the security example, another would be mining right now; over the past year, the DJ mining index is up 242.44%, outperforming all other industries. By using this method, you're portfolio would have made about those returns over the past year No need to hand-pick stocks, etc. etc. Just continue to rotate into the best performing index/industry/sector from a given date.
Grob mentioned 200 sectors he monitors, easyguru has 239 in his TC2000 database. I wonder how many sectors one show have, and how big/small a sector should be. easyguru: Any ideas which construction method for sector indices would be most appropriate? And which sector size would be the right balance between too sensible and too static? Personally, I am tending to a arithmetic, equal dollar weighted index which gives each component equal importance. Regarding the number of sectors I lean towards a number of 59, as this the number of industries as defined by the GICS (Global Industry Classification Standard) of Standard & Poor's. Thanks, agrau
There are 12 sectors: Basic Materials Capital Goods Conglomerates Consumer Non Cyclical Consumer Cyclical Energy Financial Healthcare Services Technology Transportation Utilities Michael B.
Standard & Poor's formulates this like this: "GICS consists of 10 economic sectors aggregated from 23 industry groups, 59 industries, and 122 sub-industries currently covering over 12,000 companies globally."