Impossible Tape Trader
- Thread starter grahamglover
- Start date
The IB data feed seems to be a bit too filtered for reading the flow of trades, unless this has changed in the last year or two. I initially used them when trading the Bund but their data feed into other platforms had like a one second resolution. So would see everything in the DOM and Tape update only every second (or maybe was half a second.
@Zzzz1 Looks like we weren't entirely correct - there are two independent markets for Mini Nikkei. J-Net and Regular Auction (http://www.jpx.co.jp/english/derivatives/rules/j-net/index.html). Volume is reported at the end of the trading session for both markets. Fortunately Regular Auction volume far outweighs J-Net volume. Though it is good to keep in mind that this is not a 100% centralized market. J-Net interestingly allows 'Combo Trades' - which is way to complex to attempt to understand the order flow in, since upto 6 different instruments can be excecuted in simultaneously - so for example a 100 volume transaction in J-Net could just be 1 leg of a 6 instrument spread, which really reduces the information value of that transaction. I only get data for the regular auction which accounts for 95% of volume so all is good in the world still. Started a separate thread to understand the micro-structure here.
Here is a look at the historical tick data I have to work with.. Quote is straight forward inside Bid / Ask volume. Any idea what the last value for a Trade update is? E.g. Trade,15975,5,15955.3046
Does 15955.3046 represent the value or what the?!?
@Zzzz1 Looks like we weren't entirely correct - there are two independent markets for Mini Nikkei. J-Net and Regular Auction (http://www.jpx.co.jp/english/derivatives/rules/j-net/index.html). Volume is reported at the end of the trading session for both markets. Fortunately Regular Auction volume far outweighs J-Net volume. Though it is good to keep in mind that this is not a 100% centralized market. J-Net interestingly allows 'Combo Trades' - which is way to complex to attempt to understand the order flow in, since upto 6 different instruments can be excecuted in simultaneously - so for example a 100 volume transaction in J-Net could just be 1 leg of a 6 instrument spread, which really reduces the information value of that transaction. I only get data for the regular auction which accounts for 95% of volume so all is good in the world still. Started a separate thread to understand the micro-structure here.
Here is a look at the historical tick data I have to work with.. Quote is straight forward inside Bid / Ask volume. Any idea what the last value for a Trade update is? E.g. Trade,15975,5,15955.3046
Does 15955.3046 represent the value or what the?!?
What data feed is that? Maybe vwap or some other sort of average of traded prices? By the way, how do you derive meaningful information from the tape when you don't look at the order book? I only see top of book quote updates in your Screenshot. . .
Thanks re the Nikkei update, much appreciated, Japanese exchanges change things very often, hard to keep up unless one regularly trades specific local contracts.
Thanks re the Nikkei update, much appreciated, Japanese exchanges change things very often, hard to keep up unless one regularly trades specific local contracts.
Its just quote data obtained from one of the main institutional data vendors.. file is huge and goes back years. Crazy how much more info those guys have to work with then us retail guys with retail data feeds. Pretty sure this is what quants work with.
Top of book yes - but it tells me how much was bidding and how much traded against it and vice versa for the offer. The order book / depth data I don't currently posses but trading the tape here remember. The order book is going to give additional decision support.. how thick are the bids vs offers overall, spoofing or real, wtf knows...
Think you are onto it with VWAP!
Top of book yes - but it tells me how much was bidding and how much traded against it and vice versa for the offer. The order book / depth data I don't currently posses but trading the tape here remember. The order book is going to give additional decision support.. how thick are the bids vs offers overall, spoofing or real, wtf knows...
Think you are onto it with VWAP!
Last edited:
Yep is Vwap. Finally got all the data into charting platform. Now working on some sort of historical order flow visuals to review - pretty much looking for things like @MrMuppet discussed. Sure it's historical but I hope to get a feel for how much liquidity pulls up price action, what sort of volume flow into the market preceding and during a move and bid / offer games etc. To do this I aim to highlight:
1. Short term volume accumulated at prices (usually calculated using some quick mental arithmetic of the prints or profile changes)
2. Spikes in volume (pace of the tape / aggressive prints) and associated price discovery
3. Prints above and below the inside bid / offer
4. Large market trades and if they were hitting or lifting (can highlight large market orders and / or resting orders)
5. Large Bids / Offers
6. Large iceberg orders
Any other suggestions to really disect this market?
Was thinking of deriving some Cumulative Delta calcs during certain conditions to see if that gives us some leading information. For example if volume is elevated, are participants mainly hitting or lifting? You'd think hitting would imply downwards follow through but often in other markets you see these guys usually getting trapped.
Anyway, just an update. The real goal should be time in front of the prints but since this is a fresh market to me I want to look at past behavior to see how easy it is to lean on prices, how long to allow for a pop in prices etc. This helps with the mind games when watching the real-time action.
1. Short term volume accumulated at prices (usually calculated using some quick mental arithmetic of the prints or profile changes)
2. Spikes in volume (pace of the tape / aggressive prints) and associated price discovery
3. Prints above and below the inside bid / offer
4. Large market trades and if they were hitting or lifting (can highlight large market orders and / or resting orders)
5. Large Bids / Offers
6. Large iceberg orders
Any other suggestions to really disect this market?
Was thinking of deriving some Cumulative Delta calcs during certain conditions to see if that gives us some leading information. For example if volume is elevated, are participants mainly hitting or lifting? You'd think hitting would imply downwards follow through but often in other markets you see these guys usually getting trapped.
Anyway, just an update. The real goal should be time in front of the prints but since this is a fresh market to me I want to look at past behavior to see how easy it is to lean on prices, how long to allow for a pop in prices etc. This helps with the mind games when watching the real-time action.
Why? Wouldn't this show agression?
Does not provide clear entry signals,hence completely redundant for me.
Not sure if I will get any clear entry signals out of it (or if I've ever been able to define my trading in clear signals). More like volume picked up, someone continually sticking the bid, correlated market moved up and bids stick, see some games screw over some traders then finally work a bid....
This analysis will hopefully give me a feel for how much volume tends to pull up the market, what the big players are swinging, are they refreshing orders or getting filled across prices and using sweep tactics etc.
In regards to CD - If looking back at historical order flow I would have thought would be able to identify some sort of imbalance based on the CD. Especially if you only measure how much aggressive buying vs selling is occurring at peak volume (lots of interest by weak hands and strong hands). I guess the key is that we are only looking for imbalance since the tick volume trading has different intentions (new longs vs pukes) and won't necessarily tell you to go short or long. You would think there is some sort of imbalance at reversals though... not sure if using this would disqualify me from 'Tape Trading'.
May as well get off my ass and do the analysis and then figure out if anything useful though eh?!
Last edited: