Crowd Forecasting Discussion


Do you have a question about our Crowd Forecasting feature?

Post a reply to this topic with your question...
The first official day of Crowd Forecasting is over and I've just run a quick analysis and I'm pretty excited about the findings.

So first of all the theory says that the crowd, if motivated, together will consistently beat the experts.

The only analysis that I've done so far is on the ES because that is the most traded symbol that we look at and also the one that had the most forecasts submitted for it.

The average forecast closing price for today (20 Dec 2010) was 1240.69 and the actual closing price was 1241.25. That's a 0.56 difference! That is phenomenal in my opinion. How can the average of the crowd be that accurate? Remember that everyone who submitted a forecast could see this average from when they submitted the forecast and this would have been locked at this value from 9:30am EST this morning.

Now here is the interesting part and what I am probably going to try and setup a report for. How many traders forecast better than the average crowd forecast? There was only 1 trader that did better and he/she forecast 1241.50 which was 0.25 points off and 0.26 points better than the average - i.e. just 1 tick better than the crowd.

Anyway, as you can tell I'm pretty excited about the first day's result. Hopefully tomorrow and future days will be as accurate.

More Info:
Crowd Forecasting Explained
Submit a Forecast
Today the crowd didn't perform as well as yesterday but it did perform better than most of its participants on an individual basis.

The ES closed at 1250.75 and the crowd predicted 1245.79 so the crowd was just under 5 points away from the mark. Only 4 traders managed to outperform the crowd and they missed the mark by 1.75, 1.85, 2.00 and 4.75 points.

I'll keep updating this topic with the results. Those who submitted forecasts and have their email enabled on the site will receive an email each day after trading with more information about their performance.
The crowd continues to forecast better than most of the traders again and is slightly better than yesterday.

Today the crowd forecast 1249.61 for the ES and the actual close was 1254.50 for a difference of 4.89 points. There were 3 traders who forecast closer than that and they were 1.75, 1.9, and 2.75 points off the mark. They, respectfully, outperformed the crowd by 3.14, 2.99, and 2.14 points.

I haven't run the figures yet but my guess is that the traders that are outperforming the crowd are different traders each day. I will be generating that information at a later date.

Those that have been submitting forecasts, no matter how good or bad you are, will be privy to information such as this when I produce it as a thank you for being one of the crowd.
Today's (Thursday 12/23) close in the ES was 1253.00 and the crowd's average was 1251.44 so the crowd was 1.56 points lower than the actual close.

Only 2 traders did better than the crowd and they were 0.50 and 1.00 points away which means that they beat the crowd by 1.06 and 0.56 points. Not much of a margin there.

This is the end of the first week and I'm very happy with the results of the crowd forecasting. It was a short 4 day week and next week is going to be the same and volume will probably continue to remain low until everyone comes back from the break on 3 January.

Have a great holiday weekend everyone!
Haven't had a chance to calculate the results from today. I'm going to see if I can get this sort of report automated...
Originally posted by day trading



This is the end of the first week and I'm very happy with the results of the crowd forecasting. It was a short 4 day week and next week is going to be the same and volume will probably continue to remain low until everyone comes back from the break on 3 January.

Have a great holiday weekend everyone!


So far the results look encouraging. Should be interesting to see how the forecasting goes when normal volatility returns in January.
Apologies for not updating this topic. We're working on 3 improvements:

1. To get these stats generated automatically.
2. Make it easier to forecast directly from the Daily Notes page if you're already logged in.
3. Reworking the way that elite and experts are calculated to make it more representative of accuracy of forecasts.

I'll post more details when this is available.
Originally posted by day trading

...
3. Reworking the way that elite and experts are calculated to make it more representative of accuracy of forecasts.
...

The "formula" has been reworked and expert and elite status is now just dependent on how close you were to the actual close and no longer compares how you did against your peers. After discussion with other site members and running some tests we determined that this was the fairest way to measure it. It does, however, mean that the maximum worst score of 100 no longer applies and it can now be much higher.