Two questions. Forgive me if these are painfully obvious to all the trading gods here. I'm asking these because I don't know and I have nobody else to ask. 1. Why are the charts different between Yahoo, CNBC, and MarketWatch? I was looking at SPY tonight on Yahoo, CNBC, and MarketWatch, and found the candlestick charts are not consistent between the three. All three of these are from the same day, Monday, July 1st, 2019 CNBC says they are from NYSE Arca MarketWatch, says the same, Yahoo says the same but also, 'Nasdaq Real Time Price"... Isn't this from the NYSE? Are they actually getting their numbers from different sources? Shouldn't the price of SPY be consistent regardless of the charting source? Yahoo CNBC MarketWatch Market Watch's and Yahoo's after-market long candles seem like they are the same, but they aren't. Market Watch's long down candle @ 4:13pm Yahoo @ 4:28pm Which leads me to Question 2. 2. Can somebody explain these long candles? They seem legitimate, but they don't make sense. They almost seem manipulated by a market-maker to reach very specific prices. They're end peaks seem very deliberate and many times consistent to the penny between these spikes. Many times I will see these long candles happen from extremely low volume. Almost the same volume before and after these extremely long candles. However, at these moments, they are much longer.. I've been told from somebody a long time ago they were from Dark Pools, but the prices should move the same amount from similar volume, not 30x the length from the same volume.... I used to think they were deliberate to take out all the stops from day traders, but I later learned that your stops can't be stopped out from afterhour price movements. When these happen during the day, they likely do take out people's stops, not matter how much of a micro-second they are, yes? I honestly don't know and have been just guessing and bouncing theories around for many many months.. Thank you for your patience and time to explain and answer my questions..
Almost certainly the long shadows on these cr@p charts are the results of incorrect price data input. You could double check with a broker's demo account or better a live account. But if you have a live account and you're afraid the broker will spike prices down to hit stops on long positions, you can always trail your long positions with an additional buy order well below price. Far enough below so that it won't get triggered by genuine market fluctuations but close enough so it will get triggered by a spike and you absolutely kill it when price resumes its correct position 5 minutes later.
It doesn't matter if the data is from a free source, from a charting service you've subscribed too or from your broker via live real time charts... They are will have slightlhy different data and that has been the norm since the birth of data services. You can have 5 different demos or 5 different live accounts with charts up at the exact same time...you'll see different quotes or stalls in the data for one and it will change to another having stalls minutes later I use to think it was a problem with free services but saw the same problem with subscription services, demos, real time data, brokers and so on but its an issue when the data gets converted to a visual chart. Thus, if you're just looking at price quotes in the DOM...you'll see less extreme price quotes like that in comparison to what you see on the chart itself. Reason why some traders out there refuse to use charts for their trade decisions and they exclusively make their trade decisions via the DOM. Sometimes its so obvious via those extremely long wicks on the charts and other times its not so obvious via the differences by only a few ticks. I myself caught on to it when I was doing backtesting on the same duration time and the exact same price action... Five different data providers produce different backtest results on a strategy I was testing. The results were so extreme when comparing one to the other that one showed profits, another showed losses, another showed breakeven and so on...all with the exact same start point and end point Worst, some data providers would do "corrections" of their incorrect data at a later date without any notification. Thus, you run the same tests again on the exact same duration...you'll get slighlty different backtest results at the same data source. Also, the problem is not isolated to just the price quotes (High, Low, Close, Open)...it can be seen in the volume too from one data vendor to another data vendor. One thing I do notice, this is mainly a problem on intraday data and less of a problem on higher time frames like 1hr, 2hr, daily, weekly or monthly charts. Further, the more expensive the data vendor...there are less extreme long wicks. These particular data vendors tend to advertise their data as the "most reliable" or "most clean" instead of marketing fancy indicators and extra charting features. The latter I discover in chat rooms I use to visit...dozens of different traders using different data sources, different brokers and so on...having different Open, different Close, different High and different Low for the exact same intervals. It gets funny when different traders are using the exact same trade strategy and one has a valid trade signal while the other does not have a valid trade signal. I've myself had made trade decisions on intraday data and then after I enter the trade...a few minutes later the data changes (gets corrected)...the high changes by a few ticks or the low changes by a few ticks or the close/open of the interval changes by a few ticks... The trade I was in I would have never enter with the new (corrected) info that would show its not a valid trade signal. This particular problem happens less today in comparison to 10 years ago. Further, don't be mistaken via its a problem only on candlestick charts...its the exact same problem on bar charts too but less of a problem on "line charts". Data vendors get their data from many different sources although they ALL claim to get their data directly (point A to point B) from the exchange. If you do some deep digging, you'll discover one will get their data from someone and that someone gets it from another source and that another source gets in from someone else and that someone else gets it from the exchange. Data passing thru many different companies instead of from the exchange directly to you. Then there's those that filters or corrects their own data but they often give "bad corrections" and seem to not give a damn as if they themself can not see their own charts. Simply, you need to be very careful with your trade decisions via free intraday data or cheap intraday data and understand that the most reliable data are also the most expensive data but even the most expensive data has problems too although less than the free data sources. Its life of the intraday trader and it makes me wonder if there's some truth that if you're going to trade something...make sure you live in the same city as the exchange and your data vendor is also located in the same city as the exchange in reference to something like trading Futures. You know what other traders say...do not use market orders. P.S. I can not remember the name of the data vendor but if the data in the times & sales window did not correlate with the info on the price charts...you would see a live log window of "data inaccuracies". wrbtrader
Therefore backtesting..forward testing...any kind of testing is useless and not reliable? In addition, you cannot trade what you “see” as what you see is not true? Or you can never know if it is true?
Not exactly. If all of such is done via free or cheap data...don't spend too much time debating your results (backtests or trading) with someone else that's using a different data vendor or different broker if you're using their charts when they did the exact same backtests or forward testing and they then have different reults. You know the arguments...the strategy works for someone but someone else states the strategy does not work for the exact same stock, futures or whatever. I got wind of the problem when two traders in a chat room were arguing about their backtest results on the Emini futures for the exact same year of testing & trading while using the exact same strategy. One was using Bloomberg Professional while the other trader was using free real time data. They were consistently debating while never looking at the data source as a possible reason for the difference. At first glance, I thought the isse was that one applied their trading differently than the other trader while using the same strategy. Closer inspection...they were using different data vendors and different brokers in which some of the intervals had different quotes in the high, low, close or open. Going deeper into their data...there were those strange long wicks for one trader while the other did not have such or had less of such...occurring multiple times on the intraday data...rarely on the daily charts. Simply, compare your results with someone else doing the exact same testing via a different data source or you can compare your results via different data soures (charts side by side in real time). Without the comparisons, two traders will not know they visually are seeing a different picture. Another way to look at this, if you're an intraday trader, have different data sources running at the same time for comparison to each other or you're chatting with someone in real time that's using a different data vendor than your data source so that you can questions when one see those extreme wicks or strange quotes. Summary, there are a lot of flaws in the data for intraday traders. Trading is difficult enough on its own but the issue with bad intraday data or not even knowing you have bad intraday data until you start comparing in real time...it sucks. wrbtrader
Generally these are trades that happened off the exchange and are reported later. If you are trading with tick data you will not see these trades, as such they are un-tradeable. Yahoo, and others, may "correct" their data after the fact. Not very useful. IQFeed tick data is un-corrected but there are only a few months of tick data available. Historical daily data is corrected. Still, I use and like them. DTN used to have a product call NxCore? that could give lots (years?) of tick data. I don't think they "corrected" anything. Much more expensive Good luck
Wow, my mind has exploded... Thank everybody for your information. Especially wrbtrader for so much detail and information. Frequently, on other forums where I'm an expert on the topic, I'll avoid answering because I worry about overwhelming people with information and sometimes... Man, I just don't have the time to write a novel to explain a topic in detail for some stranger on a forum who many not even read everything I've read or appreciate it. Writing such things can be a time vampire. Seriously, I so truly appreciate the time and effort everybody made to elaborate.. I would think that if your broker didn't have the cleanest data, you could get stopped out from garbage noise. That would be $#@%@!! Yes, I've seen inconsistencies with volume too. I'm glad you mentioned it as it was a question I wanted to ask too, but forgot. You would think their would be some accountability on the accuracy of the data when so much money is involved.