Shameless plug for the "bars don't matter because they're aggregations" thread.

In fact, the basis for this thread speaks so clearly -- if two charts showing identical transactions give quite different signals because (likely) the starting ticks are misaligned, or (possibly) because a few extra data aggregations took place, then it should be abundantly clear that the OHLC of those bars are completely random, and using these 4 random values for any given bar to arrive at some conclusion has zero edge.

(back to my bunker)

josh 

... In fact, the basis for this thread speaks so clearly -- if two charts showing identical transactions give quite different signals because (likely) the starting ticks are misaligned, or (possibly) because a few extra data aggregations took place, then it should be abundantly clear that the OHLC of those bars are completely random, and using these 4 random values for any given bar to arrive at some conclusion has zero edge. (back to my bunker)

I agree. The OHLC of any bar is kind of useless, and this means that candlesticks are kind of useless too. I know there's tons of ppl that like candles, but ... if they'd just shift the phase of those bars (IOW start the bars at a different place in the tick data stream), then they'd see that the candles CHANGE. So when you look at a candlestick, it's ONLY showing you what happened WITHIN that bar, you can't use a candlestick to tell you what the market is doing outside of that one single bar. That's what candlesticks were created for ... for DAILY data when ppl did not have intraday data. But now we all have computer charts and real-time data, so we should not be using candlesticks lol.

For example, look at a 1 minute (or whatever) chart... on paper make your own 2-minute candlesticks out of those 1 minute bars. Now shift the phase by 1 minute (50%), so you start on a different bar. You'll see the candlestick "patterns" are different now. LOL

I'll conclude by saying... price-based charts/bars are mathematically correct, all the other ones are not. BUT, sure, you CAN still get some usefulness out of other charts, because they take a time-window of the data and plot it, but they'll never be as good as a price-based chart (range charts or renko).

Volume charts must be better than time charts, because the market runs off volume, so you want to plot that instead of time, because the market doesn't run off time... so by using time charts you're just clouding the real data, either putting too much data or too little data into one time bar.

But range charts run off price, and PRICE IS KING. Price is what moves the markets, not volume. Sure, there's volume, but the market participants (and their volume) make all their decisions based on price. So use range charts (or renko) and don't waste your time on candlesticks, IMHO. But hey, everyone can do whatever they want, it's logically wrong, but hey, whatever makes them happy.

thoughtful 

Volume charts must be better than time charts, because the market runs off volume, so you want to plot that instead of time, because the market doesn't run off time... so by using time charts you're just clouding the real data, either putting too much data or too little data into one time bar. But range charts run off price, and PRICE IS KING. Price is what moves the markets, not volume. Sure, there's volume, but the market participants (and their volume) make all their decisions based on price. So use range charts (or renko) and don't waste your time on candlesticks, IMHO. But hey, everyone can do whatever they want, it's logically wrong, but hey, whatever makes them happy.

I'll respectfully present another viewpoint here. I'll use a little market profile parlance and cliche, but I think it's the best way to describe time, price, and volume.

Time regulates markets. That is, you can't trade on a holiday, or on a weekend, if your exchange says you can't. While they are mostly 24 hours these days, the predictably high range at the open and at the close, along with the predictably low range during the lunch hours, shows that time is indeed a major factor in markets. Whether aggregating the data by time is helpful, well that's another matter.

In principle agree about price. Your profit/loss is based on the current market, which generally hovers around the last price. But I'd have to be picky and say that "price" doesn't move anything. It is a record of a transaction, assuming you're talking about the last trade. To use MP-speak, it is the advertising mechanism of a market to attract buyers and sellers. So, this advertising may attract the interest of participants, and their capital inflows (the measure of which is shown by volume) moves a market. Technically of course, a market can move with no volume (no trades), since a market is nothing but quotes. But let's not get into that here, as I think I've already derailed the conversation far enough!

Something to keep in mind is that all charts are just a representation of market activity, supply v. demand, sentiment, emotion, etc. There is no magical chart type. I know a group of traders that swear by price type charts - range, renko, PnF, etc. But all of those charts have particular settings as well so you are creating the way you view the market you are trading. There is no pure, true view into the market that you are trading. Nothing solves this, not multiple time frames, not multiple ranges, not multiple types of charts. You are viewing market activity through a very narrow lens.

I think there are valid arguments for every type of view and many people make time based candles work. Plenty of people make range and renko work, plenty of people only trade from a DOM. I would argue that if you are following someone who trades a 2000 tick chart and they claim/appear to be consistently profitable, that same strategy would also work on a 4000 tick chart, a 10,000 tick chart, a volume chart, a 5m candle chart, etc.

I really don't think there is any magic to a particular chart type or setting. What works on one should work on another if the strategy is robust.

Just as a visual comparison load a tick chart of say 600 and plot volume below in a different frame.
Then open a chart of a 600 volume chart and plot the volume below. (notice the difference.)

Mack has mentioned that no two TICK charts will be alike. I think he brought this up in one of his YouTube videos over the past week. I hope this is helpful.

start around the 14 minute mark. The whole video has some good tidbits about choosing the right broker and getting quality data feed and internet connection.

Good day Justin

Enough good information about why tick charts are not same. It is not necessary to trade on the same bar. I beleive you trade on the patterns.

If you can elaborate more on what you tarde and how you trade, I can add more to it. You may do it thru Futures.IO or otherwise. My email is [email protected]

Raja

I use PATs and here is what I suggest: Don't worry about it and trade the price action you see.
Just yesterday, there was a long run from top to bottom where Mack missed it due to his data, whereas mine showed a good entry. Later in the day, there was a nice one where Mack's data showed a good entry, mine didn't, so I missed it. No big deal.

I agree with Rbruhn, Im also a PATS trader and use ninjatrader and my candles are not exactly same as Mack however their mostly the same. So just trade what you learned from Mack and all is good.

There was a time some years back when I compared multiple data feeds because I was working with orderflow which was based on trade volume in each tick of a bar. I had been chatting with a friend in San Jose, CA and I was in Blythe, CA on the border of Arizona and California. We were comparing charts and he suggested running ping tests to see why they were different. Comparison was between esignal(internet), tradestation(internet), IB(Interactive Brokers), BMI later DTN(on satellite) which had a sat receiver that used the RSI port and not RJ11 internet port. What an eyeopener, what a learning experience. PingPlotter couldn't be run on the satellite feed, but could on the the other three. PingPlotter is useful for seeing the actual data path from your computer to the exchange. TradeStation permitted the trader to select the server the data was from as datafeed vendors may have multiple servers around the country, or around the world. DTN satellite had one upload link for the US but it had the sat latency up and down. Beyond that each time you connect to the datafeed its internet path to get to the exchange may take a different path. Ticks get lost in the hops between your machine and the exchange. No one except computers at a co-located server have minimum loss. Sit in a trading room with a "guru" that uses tick charts and most every trader will not get the same signals, some won't even get the signal when using the same software and different machines. At that time of the testing, IB used a filter and lost 1/3 of the ticks, TradeStation bundled ticks(may not now), esignal had the shortest path to the server. The eye opener was tick loss at the internet hops, and the route the internet chooses to send the data, sometimes more hops, sometimes less. Being able to choose the TS server was an advantage, but the tick bundling was a negative. Then along came a trader/broker and Steidlmayer proof reader that swore by CQG(professional and expensive feed, not the inexpensive retail feed) and he said the software permitted him to plot all ticks OR only changed ticks. That will make a huge difference, but does eliminate noise. A hop path difference, depends on internet switching. For example one internet provider went from Blythe to Los Angeles to San Francisco bay area, hopped around the bay, then eastward across the Sierras, then more direction changes across the US and finally to Chicago. TradeStation on another ISP the path went from Blythe eastward to Phoenix, AZ to Texas then north to Chicago. It had the shortest path. Bottom line get PingPlotter and check your computer, internet hops and look for packet loss along the way. Good luck.