Data Historical Machine Learning for Trading Price Volume - - PowerPoint PPT Presentation

Machine Learning for Trading Financial Investing

Dealing with Data

Data

• Historical
• Price
• Volume

How Data is aggregated

• Many trades
• Different Exchanges
• Tick

– Finest resolution of data (minimum size)

• Price (1c)
• Upward, downward

movement (change in price)

– Cell Match

• Buy and Sell Match

9965 9921 9917 9982 9937 9917 9977 9936 9929 9959 9880 9900 9920 9940 9960 9980 10000 1/23/18 2/23/18 3/23/18 4/23/18 5/23/18 6/23/18 Price Volume: 200 100 300 100 200 100 300 100 500 100

• Buy/Sell Match (not guaranteed at specific

times)

• Different Exchanges.

9966 9916 9989 9947 9959 9929 9998 9944 9956 9997 9991 9935 9995 9961 9939 9907 9919 9996 9979 9958 9860 9880 9900 9920 9940 9960 9980 10000 10020 1/23/18 2/23/18 3/23/18 4/23/18 5/23/18 6/23/18 Price Time

How Data is aggregated

• Many trades
• Different Exchanges
• Tick

– Finest resolution of data (minimum size)

• Price (1cent)
• Upward, downward

movement (change in price)

– Cell Match

• Buy and Sell Match

Volume: 200 100 300 100 200 100 300 100 500 100

• Buy/Sell Match (not guaranteed at specific

times)

• Different Exchanges.

• Lots of Data
• Consolidate Data

– Time Epoch

• Minute by Minute
• Hour by Hour
• Daily

– Open, Close – High, Low – Volume – Combine different exchanges

Price anomaly

• Example 1:

– Blue Line: – What are the price drop here?

Price anomaly

• Example 2:

– And here, what are these price drops?

Price anomaly

q CEO Quit q Dividend Cut q Stock Split

Which makes most sense?

Price anomaly

q CEO Quit q Dividend Cut q Stock Split

• Next day 4 shares of

75, 4x75=$300.

• Next day 2 shares of

125 – 2x125=$250.

Why do Stock Splits?

• Price becomes to high: less liquid, less volume.
• Buy in groups of 100 (typical)

– 50,000 for $500/stocks.

• Exceptions/Why make an Exception?

– Berkshire Hathaway BRK.A.(doesn’t split it shares).

• Warren Buffet’s holding company.
• >$110,000 (Volume : 450 shares per day)
• Keeps the price high to deter short time traders

creating excess volume.

– Seabord SEB, $2,660 – NVR $703 – GOOG -- $618 no splits. Started at $100 in 2004

2017 prices

https://www.forbes.com/sites/investor/2011/07/25/bershire-seaboard-google-priceline/#78528b8d64ab

Why Split?

• Less Liquid, less volume
• Options on stocks are usually traded with

regard to 100 shares.

• Finely Tuned portfolio (harder with high

priced stocks).

• How to deal with

stock split data

• Not short stocks when

company value has not change:

– Adjusted Close. – Current Day:

• Adjusted close = close
• Back in history need to

adjust it.

this situation. This green line represents the actual price

• How to deal with

stock split data

• Not short stocks when

company value has not change:

– Adjusted Close. – Current Day:

• Adjusted close = close
• Back in history need to

adjust it.

this situation. This green line represents the actual price

Dividends

• Dividends,

– Annually – Quarterly

• Example

– $100.00 – $1.00 dividend announced (1%) – Stock value?

• Assume consensus
• f stock vale is

$100.00 (may not be the price).

Here we have: $1 + $100 share 1) What is the value of the stock the day before dividend payout? 2) What is the value on they day of payout?

• Day before

and day off?

• Answer:

– $101 day before – $100 day after.

Adjusting for Dividends

• Similar to stock

splits.

• Red line is

Adjusted close.

Survivor Bias

• SP 500

– Membership Change.

• Simulate todays list in the past using

historical data

– Current list survived in the past.

• Survivor Bias-

– Inflation of funds that remain when poor performers are not part of the equation.

• Need survivor bias free data (costs money).

– To more accurately measure strategies working

• n historical data.

– Need to know historical lists of companies.

https://www.elitetrader.com/et/threads/can-anyone-recommend-a-source-of-historical-data-without-survivor-bias.212720/