Market weighted Average TEV

[Bernie]

Hello,

Just a quick question on how this is compiled. Is each stock's accumulated TEV compared to its market cap to produce the percentage? From there I would imagine each stock is weighted according to their share of the total market cap and then added together to produce the final result?

Thanks,

Bernie

[Pascal]

Hello,

Just a quick question on how this is compiled. Is each stock's accumulated TEV compared to its market cap to produce the percentage? From there I would imagine each stock is weighted according to their share of the total market cap and then added together to produce the final result?

Thanks,

Bernie

Yes, you are right.

Pascal

[mklein9]

Pascal, this brings up a question that has been in the back of my mind for a while. Many components of MF analysis are market weighted, which is market cap weighted according to your answer to Bernie's question. Quite a few of the stocks in the MF universe are very, very low capitalization, such as ENER ($20 million), CSKI (under $40 million), FEED (under $30 million), etc. In total about 100 are under $1 billion, while about 300 have market caps over $10 billion and about 30 have market caps over $100 billion. Of course what this means is that when market weighting, it would seem that the MF related results are dominated by a relatively small number of companies. There is a range of over 10,000:1 in market caps.

On the other hand there seem to be a few parts of the system, like OBOS as far as I can tell, that are simply votes -- every company gets an equal vote in the outcome (or, at least, in its sector's outcome).

Of course most of the institutional funds we care about are going to/from the big cap stocks so this seems appropriate. I'm just wondering if this is an accurate portrayal.

Thanks,

-Mike

[Pascal]

Pascal, this brings up a question that has been in the back of my mind for a while. Many components of MF analysis are market weighted, which is market cap weighted according to your answer to Bernie's question. Quite a few of the stocks in the MF universe are very, very low capitalization, such as ENER ($20 million), CSKI (under $40 million), FEED (under $30 million), etc. In total about 100 are under $1 billion, while about 300 have market caps over $10 billion and about 30 have market caps over $100 billion. Of course what this means is that when market weighting, it would seem that the MF related results are dominated by a relatively small number of companies. There is a range of over 10,000:1 in market caps.

On the other hand there seem to be a few parts of the system, like OBOS as far as I can tell, that are simply votes -- every company gets an equal vote in the outcome (or, at least, in its sector's outcome).

Of course most of the institutional funds we care about are going to/from the big cap stocks so this seems appropriate. I'm just wondering if this is an accurate portrayal.

Thanks,

-Mike

Yes Mike, you are nailing here one of the key component of the system: how we are weighting things.
The most "clever" way to weight is the one used in the 20DMF, because it is the closest to what funds are doing (in my opinion.) Funds like to invest in sectors, because of their need for diversification (spread between different stocks within one sector) and the need for liquid opportunities (you just cannot put all your money in one tiny stock on one sector.)

The 20DMF is a sectors based method. Within one sector, each stock is weighted by a proprietary method that is close to a capital weighted method, but much more dynamic. However, each sector has an equal weight. This means that the computer sector (including AAPL) weights the same as the tiny solar sector. The reason is that you of course need to detect money moving in/out of the market, but you also need to see if there is sector rotation or not. The 20DMF is doing everything together and I believe that it is the key reason why it works.

I might be telling too much here about the internals of the system, but this VIT group is relatively small and since I have been enjoying many clever posts from others, I like to share more about what we are doing.

Other indicators on the web site have a different weighting method. For example, the Market Weighted TEV for 1000 stocks is my vision of the "whole money regarding the whole market." TEV combines both the large and small players and 1000 stocks is what everyone is doing. It is not very clever: it is the bulldozer type of indicator.

The number of days indicator is similar to the OB/OS of the 20DMF, but much "safer" to use: it is safe, because it does not issue many signals, but the signal quality is better. However, the returns are lower, because it forces to stay in Cash for too long (for example, as of yesterday's close, we were still two days from a BUY signal, while the 20DMF issued its buy signal many days ago.)




Pascal

[Bernie]

Thank you Pascal for your responses (and Mike for the further question) , I did not realise that the sectors were evenly weighted for the 20DMF. I have been doing some EV calculations for the Australian market and as Mike mentions I found that the Large player strength of the ASX 20 looks the same as the Large player strength of the ASX 200 due to the market weightings based on each stock's market cap. Perhaps if I applied an equal weighting to each of the GICS Sectors and/or industry groups I would come up with a somewhat different picture.

Within this ASX 20 its pretty much Financials and Metals and Mining that make up the most of this. A very small market.

For what its worth I thought I might share the graphs of what I have come up with. The data is calculated minute by minute but I have only used the end of day values for the graphs. One is for the ASX 20 (XTL), the other for the Metals and Mining index (XMM). The dates only go as far as mid August for XTL and late September for XMM due to the time required to perform these calculations (as they rely on daily Large Player EV calculations for the required stocks, and my code is slow to do this :p). Also the 10 day average on the XTL graph up to the 4th Oct is not correct.

I think the Metals and Mining index looks a little similar to the Coal Sector LP strength for the same time period?

11802
11803