S&P 500 cash weekly competition for 2013

It's interesting that the people with a negative bias, which is clearly a lot of the group, tend to predict bigger moves each week.
 
Just to add twopenneth - the weighted average favours those that are performing well and marginalises those that are performing badly over a long run average. It is not likely to respond materially with an outlier prediction unless that outlier comes from one of those performing well. I'm trying to say it self polices Pat and there isn't a need to be concerned about it. With the Mean, your assertion is certainly true but for the Weighted Average, the existence of an outlier as spoiler is very, very limited.

On another note, I have observed that Pete does v.well during bearish markets (as he is a perma-bear) and it is no surprise to me that Viel has done well as he is a perma-bull and conditions right now are bullish.

Over the weekend, I might take the 2012 results to see where in the bull/bear spectrum regular participants are in their predictions vs actual. Then I'll draw it up as a distribution which will be quite 'blocky' but probably quite entertaining to see.


It should not be that difficult to try out next quarter. That is keeping the competition as is and have a second Weighted Average with top and bottom outliers removed. I'll call it WA2.

In order not to bias the mean average calculation, it will be
entered last after WA1 and will not be included in calculations. So competition rules as before.

I will also be adding two columns for calculating PIPS made + Running Total PIPS alongside medals table. It is easy enough to calculate using excel so should add some P&L to the comp.

Once again really pleased with the development of the WA and keep it up everyone. (y)
 
Quick question1
Shall I created new spread-sheet for QTR2 or just add columns to existing spread-sheet?

Does anyone have any preferences - bearing in mind any future calculations or stats you may want extract?


Quick Question 2
For consistency - I think we should carry the WTAverage Sum & WTAverage Points forward?

This will also provide running total in WTAverage Points.


Any ideas, ayes or nayes?
 
Last edited:
One final one - the Sum of the difference over the year does not predict annual performance in comp either.

Note for this, I have applied squareroot of the sum to scale it back to position. Log natural was not granular enough.

oimg


https://docs.google.com/spreadsheet/ccc?key=0AnFF2Rblu36wdGRNQjNXTzVjdldJb2dXdjRVTGs5VVE#gid=10
 
Quick question; shall I created new spread-sheet for QTR2 or just add columns to existing spread-sheet?

Does anyone have any preferences - bearing in mind any future calculations or stats you may want extract?

It is far far easier for me to do this if you leave all the results in a single sheet for the year like we had for 2012
 
It's interesting that the people with a negative bias, which is clearly a lot of the group, tend to predict bigger moves each week.


Looking at the difference between Estimated and Actual ATRs; the average ratio is approx 2:1 respectively.
 
It is far far easier for me to do this if you leave all the results in a single sheet for the year like we had for 2012

Ok will do (y)

I've added 2nd quick question too :cheesy:

My feeling is we should maintain weighted metrics.
 
Quick Question 2
For consistency - I think we should carry the WTAverage Sum & WTAverage Points forward?

This will also provide running total in WTAverage Points.


Any ideas, ayes or nayes?

I say nay as the rules of the comp being applied to the other participants will differ. Keep it the same. Other derivations of 'indicators' should have the same comp rules applied regardless of how you derive them.
 
Also interesting that there was 32 up weeks last year (61.5% positive), yet 7 of the top 10 players were negatively biased. But I'm assuming got the additional podium points regularly.

I wonder if there is something anomalous about the Up/Down marking where your predictions can be very close but if they are on the wrong side of the 'direction' you are penalised. If this is the case and it was a bullish market, calling under would have been a safe strategy for regularly clocking up direction points.

I think the top scorer scored on average about 1pt per week - that would be a direction point average over the year without podiums.
 
It should not be that difficult to try out next quarter. That is keeping the competition as is and have a second Weighted Average with top and bottom outliers removed. I'll call it WA2.

In order not to bias the mean average calculation, it will be
entered last after WA1 and will not be included in calculations. So competition rules as before.

I will also be adding two columns for calculating PIPS made + Running Total PIPS alongside medals table. It is easy enough to calculate using excel so should add some P&L to the comp.

Once again really pleased with the development of the WA and keep it up everyone. (y)

Might be interesting or too similiar to the original. We shall see. Are there other mathematical formulae that could be applied ?
 
Maybe that's a rule change for consideration in future, in that the three closest predictions regardless of direction should get the podium points. But the directional call should be an additional point and not affect the podium places? So you've got 1, 2 and 3 points for three closest picks, and then everyone that gets the direction right gets and 1 additional point.

On the rules subject, I also think we should make all predictions unique, so no one has the same one.
 
Maybe that's a rule change for consideration in future, in that the three closest predictions regardless of direction should get the podium points. But the directional call should be an additional point and not affect the podium places? So you've got 1, 2 and 3 points for three closest picks, and then everyone that gets the direction right gets and 1 additional point.

On the rules subject, I also think we should make all predictions unique, so no one has the same one.

I think that makes good sense and would like to introduce into Q2. Any objectors?

Re: predictions also - first come first get!
 
Currently, the rules favour the bears because it is a bullish market and the bias to undershoot a rising market whilst still predicting upwards generates directional points. The converse would be true in a bearish market and the bulls would start to dominate. Think about the effects of changing the rules as there may be an element of short termism about the rule changes. I'm not saying don't, just consider it carefully.
 
Last edited:
Yeah, I'm not sure if it should be an immediate change as this year is already underway. Which is why I said consideration in future, i.e next years comp, but I definitely think the unique predictions suggestion should be started, as I do think that skews the overall results.
 
Using the correlation function on the datasets demonstrates the lack of correlation between performance (as measured by avg diff or sum of diff from actual close) and position in comp. Avg Diff vs Position is not correlated (0.01), Sum of Diff vs Position is weakly correlated (0.27)

I'm going to try being bearish for a while rather than accurate - it seems you get rewarded for using a blunt tool.
 
Last edited:
Currently, the rules favour the bears because it is a bullish market and the bias to undershoot a rising market whilst still predicting upwards generates directional points. The converse would be true in a bearish market and the bulls would start to dominate. Think about the effects of changing the rules as there may be an element of short termism about the rule changes. I'm not saying don't, just consider it carefully.

I think the podium should go the nearest forecasts either way.

eg: Week ending 22nd Feb, I got bronze + 1 point for being -17.60 ticks away from Friday close whilst Robster (4.40) and two others got nothing despite being closer because they had direction wrong. That is 4x distance.

It should be about proximity to net result not direction imo and I'm still in favour.

Podium points only goes to nearest forecast in any direction. +1 Point for correct direction?

Also, as we are at the start of a new Qtr2 - so it is an opportune moment.

Anybody else like to vote?
 
US consumers shrugged off rising taxes in February as they earned and spent more, in another sign the world's largest economy is on the road to recovery.

The US Commerce Department said Friday that consumer spending rose 0.7pc in February from January. It was the biggest gain in five months and followed a revised 0.4pc rise in January, which was double the initial estimate.

A sustained pace of steady job gains is starting to boost wages, which should help to provide some cushion for households from higher taxes and support economic growth.

Americans' incomes rose 1.1pc last month, following January's 3.7pc plunge and December's 2.6pc surge. The huge swings reflected a rush to pay bonuses and dividends in December before taxes increased.

After-tax income also increased 1.1pc last month. That allowed consumers to put a little more away. The saving rate increased to 2.6pc of after-tax income, up from 2.2pc in January.

The gains in spending and income follow other signs of an economy gathering momentum. Hiring is up, businesses are spending more, the stock market is hitting record levels and the housing recovery is strengthening.

More spending by consumers should boost economic growth in the January-March quarter after a lull at the end of last year. Consumer spending accounts for 70pc of economic activity.

Jennifer Lee, senior economist at BMO Capital Markets, said the increases suggest consumer spending could be growing in the first quarter at an annual rate of more than 3pc. That would be the fastest gain in more than three years and more than double the 1.3pc rate in the fourth quarter.

"More encouraging news on the U.S. economy," Lee said in a note to clients. "Always welcomed but especially now with weakness elsewhere in the world."

Most economists predict the economy is growing at an annual rate of roughly 2.5pc in the January-March quarter. That would be a vast improvement from the 0.4pc growth rate in the October-December quarter, which was held back by slower company stockpiling and the sharpest defense cuts in 40 years.

Inflation, as measured by a gauge tied to consumer spending, increased 1.3pc in February compared with a year ago. That's well below the Federal Reserve's 2pc target, giving the central bank room to keep stimulating the economy without having to worry about price pressures.

Consumers spent more at the start of the year even after paying higher taxes. An increase in Social Security taxes has reduced take-home pay for nearly all Americans receiving a paycheck. And income taxes have risen on the highest earners. The tax increases both took effect on Jan. 1.

One reason the tax increases haven't slowed the economy is companies have accelerated hiring and are slowly but steadily increasing wages.

Employers have added an average of 200,000 jobs a month since November (Xetra: A0Z24E - news) . That helped lowered the unemployment rate in February to a four-year low of 7.7pc. Economists expect similar strong job gains in March.

Businesses are also investing more in equipment and machinery, which has given factories a lift after a disappointing 2012.

And the housing recovery that began last year appears to be sustainable. In February, sales of previously occupied homes rose to the highest level in more than three years. The gains have helped lift home prices, which have made Americans feel wealthier.

Stock prices have also surged. On Thursday, the Standard & Poor's 500 index closed at a record high of 1,569. That surpassed the previous record of 1,565 set in October 2007, a year before the peak of the financial crisis.

Three weeks ago, the Dow Jones (DJI: ^DJI - news) industrial average beat its 2007 record.
 
@Atilla - I have added Diff columns (which are now hidden) in the 2013 sheet

If you change the rules we can see a before and after although I think a quarter is not enough data for the sample.
 
I think the podium should go the nearest forecasts either way.

eg: Week ending 22nd Feb, I got bronze + 1 point for being -17.60 ticks away from Friday close whilst Robster (4.40) and two others got nothing despite being closer because they had direction wrong. That is 4x distance.

It should be about proximity to net result not direction imo and I'm still in favour.

Podium points only goes to nearest forecast in any direction. +1 Point for correct direction?

Also, as we are at the start of a new Qtr2 - so it is an opportune moment.

Anybody else like to vote?

I used to argue against but am happy to give it a go.
 
Top