Using Small and Large Datasets to Improve Your Trading

Day Trading Statistics - Large Dataset vs Small Dataset

Statistics for Trading - Small and Large Datasets

I would like to share my opinion on small and large datasets.  Are there and benefits of a small dataset?  When day trading using statistics, how can a trader use small datasets and large datasets?  Many may completely discount the usefulness a small datasets.  It is true that a large dataset is more accurate over a large period of time.  If you were going to build automatic trading systems, you would likely want a pretty large dataset, especially if you are going to do any statistical modeling.

Most of us understand the benefits of large datasets.  What about a smaller dataset, though?  Are they of any value?  I publish statistics as small as 1 week’s worth of data, not very useful to a statistician.  What about for a day trader?

What have you done for me lately?

There is one good reason for using a small dataset, as long as it is very recent.  A small dataset tells you what has been happening lately, that’s it!  This is invaluable to a very short-term day trader.  Markets consolidate, react to news and are susceptible to current world conditions.  Knowing what has been happening lately helps a day trader make trading decisions.  It may even make a trader decide to stop trading a particular market or a trader may determine to trade smaller or even larger depending on how the market has been behaving lately.

Examples of Small Datasets

At the writing of this blog post, Crude Oil has been tagging the ONR 100% of the time for the last 7, 15, and 30 days.  That is good to know.  Is that something that I’m going to base a trading system on?  NO WAY!  Is that something I want to know when making decisions on certain individual trades?  ABSOLUTELY!

Another example is IBP 1.5 tagging.  Lately, Crude Oil has only been tagging IBP 1.5 80% of the time in the last 7 days and 63.6% of the time in the last 30 days.  What is this telling me on a short-term basis?  It tells me that Crude Oil is likely to make its most significant moves in the first 2 30-minute periods.   I could check to see if this is true by analyzing the 30-Minute Range Statistics Reports that I publish.

Please don’t use the previous statistics to make any trading decisions.  Please subscribe to my statistics and perform your own analysis.

The need for Large Datasets

There is also a need for a larger dataset.  I’ve already mentioned developing automatic trading systems as one example.  This site, though, is not designed to provide that kind of information for systems developers.  I publish 1-year datasets so that day traders can compare what is happening lately to what tends to happen on a longer-term basis.  Too large of a dataset can be counterproductive for a very short-term day trader.  Why?  Because market personalities change.  What the median for a given 30-minute range over a 3 or 5-year period is not as useful to a futures day trader compared to what has been happening lately.

Having a larger dataset, say 90, 180 or 365 days worth of data is very valuable.  It can help a day trader to know what is normal and at the same time how a given market is trending.  Having a larger dataset gives you a bigger picture.

Clearly, both small and large datasets are useful to day traders.  Fortunately you don’t have to choose between the two.  Here at statsfortrading.com we provide both.  Each day trader can determine which statistics are most useful for his or her style of trading.

I hope this blog post helps you understand the benefits of using both  small and large datasets.  Please check out our YouTube Channel for videos that can help you build a statistics mindset.  Here is a link to my YouTube channel:  Click Here