|
|
|
|
|
by seanhunter
1810 days ago
|
|
One of the fundamental ideas the parent mentioned is divident/split/"corporate event" data. If you're serious about understanding equities, you have to account for these or everything will be off. If you're new to stocks, the main corporate events are:
1)Dividends - stocks sometimes pay out some cash or cash-equivalent stock to holders. This causes the price to briefly drop after the date because if you tihnk about on one day you were going to own (say) Microsoft with $100 in the bank, the next day if you're a holder, you own Microsoft (with only $95 in the bank because it's paid out a dividend) and you get $5 dividend. But if you buy on that second day, you only get Microsoft with $95 in the bank. So the price on the second day should be lower.
2)Splits/reverse splits - Companies sometimes decide to adjust the number of shares issued and the price will adjust pro rata. So for example, say the price is $100 and there are 100 shares, then next day the stock splits 2 for 1 so there are 200 shares now. The company is the same, so the price will be half (ie $50). If your market data does not adjust for these (or allow you to adjust), then these effects will wreck any kind of model you try to build. Other things you definitely need to be able to do/get:
1)Basket/index/etf compositions. If you're doing any kind of trading of stock etfs, index options/futures etc you need to know what's in the basket and how that changes. This also matters if you're trading things like the russel rebalance, where there is a lot of trading around what stocks go into or out of the indices
2)Calendars for earnings, market holidays etc. You can get this from a third-party source but you definitely need it from somewhere. |
|