ES Model

I have named this model the ‘ES Model’ as it is based on ideas picked up on the Evil Speculator blog. Mole said the following ‘Throughout the 2008 downturn confirmation of an Intermediate or higher degree top has been preceded by an upswing or at least flattening of the BAA-TYX yield spread.’ Searching around for more on this I saw that one of the contributors in the comments section at another point suggested that TLT:LQD and the IEF:LQD ratios could be used as alternatives. So I decided to graph those two ratios against the S&P 500.

Looking at the graph I formed an impression of approximately what happens before an intermediate top.

The next challange was to turn this observation into something that can be identified on an excel model. It took me a while because it sometimes takes time before my head can work out how shapes on a graph can be identified with numbers and formulas. But I now have a basic model.

It is very close to a sell signal now but has not actually produced one throughout the course of this rally since March. Looking back it was making a big song and dance around the 2007 tops and in 2008 identified the tops in May and August as well as one of the December 2008 tops. In fact the last time it produced a sell signal was 8th Decmber 2008. It came very close to a sell signal on the last two days of June and the first day of July – and that moment was followed by a drop of over 40 points on the S&P 500 which was immediately before this latest surge.

The only two signals it produced that I was not very happy with were on 19th September 2007 and 3rd October 2007. However, if you are looking for approximately where an intermediate high might be these are not that bad.

The dates since 2007 that it triggered sell signals are as follows on the table below:

 

Graph Date	S&P 500 Close	ES Model Triggers SELL when TRUE.
08 December 2008	909.70	TRUE
28 August 2008	1,300.68	TRUE
19 May 2008	1,426.63	TRUE
16 May 2008	1,425.35	TRUE
15 May 2008	1,423.57	TRUE
31 October 2007	1,549.38	TRUE
03 October 2007	1,539.59	TRUE
19 September 2007	1,529.03	TRUE
23 July 2007	1,541.57	TRUE
19 July 2007	1,553.08	TRUE
18 July 2007	1,546.17	TRUE
17 July 2007	1,549.37	TRUE
12 July 2007	1,547.70	TRUE
11 July 2007	1,518.76	TRUE
09 July 2007	1,531.85	TRUE
06 July 2007	1,530.44	TRUE
05 July 2007	1,525.40	TRUE

 

Even if it does generate a sell signal soon I would not rely on that for two reasons. (i) I have not tested how it behaves in a bull market and (ii) I’d rather use this as a work-in-progress and blend it in with other ingredients like volume and implied volatility in order to get something that fills me with a greater degree of confidence.

However, it is interesting.

 

VXO Strength Continues

The VXO continues to increase relative to the market – particularly in comparison with recent months (see my post on 10th August). Yesterday the S&P 500 dropped 1.27% and the VXO rose by 4.19%. This is a ratio of (3.31). Over the last four days the market has dropped by 0.83% and the VXO has risen by 3.76%, a ratio of (4.50), see the table below. [By the way - if it is not clear, that final column heading is supposed to say '4 Day Ratio'].

 

Date	S&P 500 Close	VXO Close	S&P 4 Day Change	VXO 4 Day Change	4 Day Ratio
11 August 2009	994.35	24.85	(0.83)%	3.76%	(4.50)
10 August 2009	1,007.10	23.85	0.14%	(1.45)%	(10.03)
07 August 2009	1,010.48	24.65	0.78%	(0.84)%	(1.08)
06 August 2009	992.49	25.44	0.51%	(0.59)%	(1.16)
05 August 2009	1,002.72	23.95	1.62%	(3.04)%	(1.88)
04 August 2009	1,005.65	24.20	3.13%	(2.85)%	(0.91)
03 August 2009	1,002.63	24.86	2.35%	3.67%	1.56
31 July 2009	987.48	25.59	0.54%	8.80%	16.31
30 July 2009	986.75	24.70	0.76%	7.11%	9.30
29 July 2009	975.15	24.91	(0.12)%	3.49%	(29.89)
28 July 2009	979.62	23.98	2.68%	2.17%	0.81
27 July 2009	982.18	23.52	2.89%	(0.34)%	(0.12)
24 July 2009	979.26	23.06	2.96%	(2.70)%	(0.91)
23 July 2009	976.29	24.07	3.82%	(3.22)%	(0.84)
22 July 2009	954.07	23.47	1.42%	(8.10)%	(5.72)
21 July 2009	954.58	23.60	2.35%	(9.47)%	(4.04)
20 July 2009	951.13	23.70	5.00%	(9.95)%	(1.99)
17 July 2009	940.38	24.87	4.36%	(7.68)%	(1.76)
16 July 2009	940.74	25.54	7.01%	(11.66)%	(1.66)
15 July 2009	932.68	26.07	5.66%	(11.36)%	(2.00)
14 July 2009	905.84	26.32	2.99%	(14.90)%	(4.99)
13 July 2009	901.05	26.94	2.27%	(11.15)%	(4.91)
10 July 2009	879.13	28.91	(2.18)%	3.96%	(1.81)
09 July 2009	882.68	29.41	(1.53)%	7.14%	(4.66)
08 July 2009	879.56	30.93	(4.74)%	22.88%	(4.83)
07 July 2009	881.03	30.32	(4.17)%	22.31%	(5.36)
06 July 2009	898.72	27.81	(3.07)%	17.09%	(5.56)
02 July 2009	896.42	27.45	(2.45)%	8.28%	(3.39)
01 July 2009	923.33	25.17	0.33%	(2.71)%	(8.11)
30 June 2009	919.32	24.79	2.04%	(13.02)%	(6.38)
29 June 2009	927.23	23.75	3.59%	(19.00)%	(5.29)
26 June 2009	918.90	25.35	2.90%	(15.53)%	(5.36)
25 June 2009	920.26	25.87	(0.11)%	(2.52)%	23.98

 

If you look back to the 25th June you will notice that over the previous four days the market had dropped by 0.11% but over the same time period the VXO had dropped by 2.5%. A clear difference to now. The VXO Elasticity – if that is the right term – is changing.

 

The table below looks further out just to illustrate the same point but over a much longer time frame. The 200 trading days prior to today was 23rd October 2008. Obviously there was a lot of fear in the system then but it’s interesting to see that since then the market is up 9.5% and the VXO is down over 60%. There has to come a point where the VXO has less room to fall relative to market increases.

 

Date	S&P 500 Close	VXO Close	200 Day S&P Change	200 Day VXO Change	200 Day Ratio
11 August 2009	994.35	24.85	9.50%	(63.75)%	(6.71)
10 August 2009	1,007.10	23.85	12.30%	(66.12)%	(5.38)
07 August 2009	1,010.48	24.65	5.80%	(55.81)%	(9.62)
06 August 2009	992.49	25.44	0.72%	(51.75)%	(71.92)
05 August 2009	1,002.72	23.95	6.61%	(66.30)%	(10.03)
04 August 2009	1,005.65	24.20	6.26%	(65.60)%	(10.48)
03 August 2009	1,002.63	24.86	10.44%	(63.63)%	(6.09)
31 July 2009	987.48	25.59	(1.06)%	(56.22)%	53.28
30 July 2009	986.75	24.70	(1.65)%	(59.02)%	35.68
29 July 2009	975.15	24.91	8.44%	(71.03)%	(8.41)
28 July 2009	979.62	23.98	7.66%	(68.43)%	(8.93)
27 July 2009	982.18	23.52	(0.28)%	(65.42)%	233.45
24 July 2009	979.26	23.06	(1.70)%	(63.43)%	37.24

 

My models are in No Man’s Land

I have a smorgasboard of models which look at volatility, volume and yields from a number of angles.

The big mistake I have made recently is to trust the clear intermediate term sell signals while at the same time not paying enough attention to the fact that strong short term sell signals were leading to weak sell offs at best. Any departure from strong sell signals seemed to be accompanied by surges to the upside on the markets.

I don’t know which model to pull results from today. Taken as a whole they are at levels which in the past have been consistent with intermediate tops. However, compared with the recent past they are not as strong as on those days when there were pullbacks.

So rather than jump to a definative forecast for market direction today I’d rather wait and see whether the market has become more responsive to the sell signals.

 

Does the recent volume increase indicate a new bull market?

I spent quite a lot of time recently looking at the differences between the start of the bull market in 2003 and the recent stock market advances since March 2009. I concluded that a missing link in my analysis was volume. Below I have some of the results from my model. First some explanation of the table.

The first row (header):

This says ‘Days Up’ and next to it is a number. The number refers to the number of trading days since the low.

 

The first column (excluding the header):

This refers to the year and possibly year and month. e.g. 62 is 1962 and 01_04 is 2001 (April). The letter after the date refers to whether this turned out to be a bear market RALLY ‘R’ or a new BULL market ‘B’. The solitary ‘C’ stands for current i.e. 09C – this is because at the time of building the model I was seeking to work out whether the 2009 stock market rise looked more like a bear market rally or a new bull market.

 

The second column:

This refers to the increase in the S&P 500 vs the low (after the number of trading days indicated by the number in the top row).

 

The 3rd column:

This refers to the 10 Day VOLUME Simple Moving Average and the difference in % terms between the 10 D SMA x days after the low vs the 10 D SMA on the day of the low.

 

As I hope you can see new bull markets are characterised by increasing volume (vs the low) while bear market rallies tend to have falling volume as the market index rises from it’s low.

 

What is a rally and what is a new bull market was determined by my visual observation of a chart. It is therefore subjective. However, I did this before working on the data.

 

I have chosen to take as the example on the chart a data point which is 20 days out from the low. That is because we are now 20 days out from the ‘low’ close of 13th July.

 

Days Up	20
Trend	% increase in Index	10 D SMA
62R	7.91%	(36.4)%
70R	7.89%	(28.9)%
73R	6.20%	39.0%
81R	5.17%	(8.9)%
82R	6.88%	(19.4)%
84R	1.56%	(5.1)%
01_04R	13.17%	(9.7)%
01_09R	8.54%	(18.2)%
02R	18.92%	(12.2)%
08_03R	7.24%	(6.2)%
08_11R	18.00%	(2.8)%
62B	13.68%	25.3%
66B	10.40%	(5.6)%
70_07RB	8.37%	6.1%
70_08B	10.38%	58.8%
74B	9.24%	22.7%
78B	2.26%	7.3%
82B	18.11%	54.9%
84B	13.54%	30.3%
98B	9.58%	3.0%
03B	9.69%	3.4%
09C	23.49%	(15.4)%

 

The 09C refers to 20 days after the March 2009 low (I am using close (not intra day) data only).

 

In all but one of the bear market rallies the volume was down after 20 days (the exception was 1973). In all but one of the new bull markets (I am excluding March 2009) volume was up 20 days after the market low.

 

On my model I can experiment and change the number of days out to my hearts content and I have a useful little scatter graph that illustrates the results. Playing around with that makes me all the more sure that the general rule I have identified seems to ALMOST always work. Certainly no new bull markets seem to start on STEEPLY falling volume unless March 2009 is the first example. The 1973 rally did however see a big spike in volume.

 

Now today the Volume 10 Day Simple Moving Average is higher than it was on 10th July when the market closed at 879. Is this a clue that we may have a new Bull Market?

 

If so it contradicts most of what we have seen since 9th March because most of the time since March, volume has been falling quite steeply.

 

Quick Post on Volume

One of the big differences between the new bull market in 2003 and recent events has been the behaviour of volume.

I have studied the difference in volume between new bull markets and bear market rallies since 1962. My conclusion is that new bull markets begin with rising volume and bear market rallies are characterised by falling volume. I think that new bull markets typically see decling volume going into the ultimate low (despair) followed by increasing volume.

This is one reason why since March I have viewed what is going on now as a bear market rally (volume rose into the low and fell afterwards). However I have noticed that going into the 10th July mini low volume (as measured by the 10 Day MA was falling but since then has gone up…).

It still looks a lot less than at the March low but I think that this is worth watching.

I am not yet sure how best to put my excel charts into a blog post. I am sure I will work it all out if I become a more experienced blogger. However, I do have in a table form an analysis of all the new bull markets and bear market rallies since 1962. This I found to be quite persuasive. I hope to put that up at some point.

Of course – I understand that these volume relationships have been studied by many times in the past. I like to have my own models and studies however.

 

VXO Elasticity v S&P 500

I am sure that the following will be too obscure for most people’s interest. It is part of a table showing the relationship between changes in the S&P 500 and the VXO. Why not vs the VIX? Well, I do think that each indicator tells something different which is why I look at lot of different bits of data and how they are changing and how their relationships with each other are changing.

As you can hopefully see from the table below there has been a little bit of a change in the relationship between the VXO and the S&P 500. For several months now the VXO has been declining at a proportionately quicker rate than the S&P 500 has been rising. That is quite clear from a wider and longer table than the one I have been able to fit in below. During that time the elasticity has waxed and waned.

However, you can see below that in the most recent days the VXO has become more responsive on the upside. Look back to 22nd July and you will see there that from then looking back four days the S&P 500 had risen 1.42% and over the same period the VXO had dropped 8.1%. By the end of last week however the S&P 500 over the previous four days had risen by 0.78% but the VXO had only dropped by 0.84% over the same period.

 

Date	VXO Close	S&P 4 Day Change	VXO 4 Day Change	4 Day Ratio
07 August 2009	24.65	0.78%	(0.84)%	(1.08)
06 August 2009	25.44	0.51%	(0.59)%	(1.16)
05 August 2009	23.95	1.62%	(3.04)%	(1.88)
04 August 2009	24.20	3.13%	(2.85)%	(0.91)
03 August 2009	24.86	2.35%	3.67%	1.56
31 July 2009	25.59	0.54%	8.80%	16.31
30 July 2009	24.70	0.76%	7.11%	9.30
29 July 2009	24.91	(0.12)%	3.49%	(29.89)
28 July 2009	23.98	2.68%	2.17%	0.81
27 July 2009	23.52	2.89%	(0.34)%	(0.12)
24 July 2009	23.06	2.96%	(2.70)%	(0.91)
23 July 2009	24.07	3.82%	(3.22)%	(0.84)
22 July 2009	23.47	1.42%	(8.10)%	(5.72)
21 July 2009	23.60	2.35%	(9.47)%	(4.04)
20 July 2009	23.70	5.00%	(9.95)%	(1.99)
17 July 2009	24.87	4.36%	(7.68)%	(1.76)
16 July 2009	25.54	7.01%	(11.66)%	(1.66)
15 July 2009	26.07	5.66%	(11.36)%	(2.00)
14 July 2009	26.32	2.99%	(14.90)%	(4.99)
13 July 2009	26.94	2.27%	(11.15)%	(4.91)

 

 

How should this be interpreted? Well, it is one of a number of things that I monitor and use in the building of my models. I have noticed that I often interpret the meaning of changes in the VIX quite differently to the way others do. I think this is partly to do with the fact that I think that the way these things should be viewed is affected by wider market conditions. Also, some people are looking for trends while others are looking for turning points.

 

Introduction

I use excel to build financial market models and I want to start sharing some of this work more widely. I believe that some of the work that I have done provides quite powerful models for navigating the ups and downs in the stockmarket – in particular the S&P 500.

Trading the markets however requires far more than a good model. It requires discipline and good management of money and risk. My models cannot give you self discipline. Some of the work I have done may be used to help with the management of risk in so far as in some of my models I use historical data not only to find good sell or buy points but also to calculate the stop loss which would have provided the most profitable trades as well as profit targets and trade closes based on the movement of a particular indicator. However, this has not been the focus of my recent work and I do think that I need to make the management of the trades including stop losses and trade exit signals a larger part of what I do again.

As you can guess from the name of this blog the main focus of my work has been on volume, volatility and yields. I don’t trust myself to look at a chart and come to much more than a highly subjective point of view over what it indicates. I do make my own charts in excel however and use that as a starting point for looking at what might be predictive. I then test what I think I see on the charts with historical data.

I am also interested in the possibility of putting other people’s theories and ideas to the test.

I don’t know how this blog will go although I’ve done one or two experiments with a blog. One of the challenges for me will be to summarise on a table the key points from my models and I am not too sure how that will go.

I have the feeling that some of the models that I have created are powerful enough for me to not want to give away the secrets of the nuts and bolts of their calculations. Whether I am right to think like that I do not know.

Some of the blogs I like are Vix and More, Quantifiable Edges, Cobra’s Market View and Frank’s ‘Trading the Odds’. These are a great source of ideas for me – so thanks guys for that.