BlowFish

That's interesting, I guess by your turn of phrase (data runs) these are histories rather than real time? I wonder what exactly do you mean by 'CME data runs'.... if you don't mind me being nosey that is. Are full order book changes available or just best bid best ask? Are order book changes/best bid best ask changes time stamped or do you need to rely on sequencing (the order of changes)? Or maybe I have it wrong and you have a leased line connection to the CME and are using a comparison to data you are collecting in in real time directly from Globex? Edit: Incidentally FT, as your testing on Zenfire, IQ.Feed and CME data runs have shown no discrepancies I wonder if the issues is actually at O66's end after all:). Quite understandable that the protocol to deliver real time quotes might drop data between zen servers and the client? I understand its fairly well compressed so loosing a packet might actually loose a chunk of data. I would guess it uses UDP rather than TCP.

Just remembered....I heard a rumour that zenfire was going to be offering a days history (on the multicharts forums) this would suggest they are introducing some sort of ticker plant. Sounds like its kind of new TS SUPPORT :: View topic - FYI: Rithmic (Zen-fire) provides historical data ...

I was wondering how you can verify it is complete and synchronised? Against what? You would need co-located ticker plant and even then data loss could be encountered at the exchange (do CME guarentee too provide every bid ask change synchronously with each trade)? Just being my usual sceptical/devils advocate self and hopefully encouraging people to think for themselves. (Which O66 the OP is clearly doing!) Tradestation also has some weaknesses in the client (even if the data was up to snuff), I wouldn't go there personally. Also just to add a comment on Multicharts too. There are some issues in the client that have been discussed recently on the TSSupport forum. Possibly fixed in version 6 beta 1 possibly not. A while back I posted some code for doing delta type calculations and various types of block size filtering. A couple of people commented that they where getting inconsistent results on different charts running in the same workspace...... well it turns out that MC can miss ticks as well as not handling bid ask changes synchronously. Anyway sadly it seems that currently we don't have too many options for this type of work though NeoTicker should easily be up to the task as well as IRT/MD at least giving some choice.

Thanks for the update. The priorities for a feed that has an order entry engine are somewhat different than those for a simple data feed. Timeliness is more important than completeness. In the old days (again not verified for years) Globex had a kind of two tier setup where clearing house/brokerage type operations would sometimes get more current pricing information at the expense of not getting every tick. Incidentally do we know for sure that bid ask and last changes are fully syncronised for DTN.IQ ? Can we even be sure that the exchange actually provides that information synchronously? (Rather than giving precedent to actual completed order information over quote changes for example). At the end of the day things might be 'good enough' depending on what your objectives are.

Not all indicators are based solely on price data. The ones that tend to still tempt me towards the dark side are ones that are not. For example market delta type indicators or breadth and depth indicators (tick adv/decl etc.) or even statistical type indicators (like VWAP & its standard deviations or volume profiles). Of course as Tams points out if your are aiming towards automation, data pre processing might well make the task simpler. As an aside anyone who is interested in automation (or a purely mechanical approach) based entirely on price action should consider a good look at Dunigans' "New Blueprints for Gains in Stocks & Grains" and the "One-Way Formula for Trading in Stocks & Commodities" (available fairly inexpensively as a single publication). Actually it's worthwhile even if simply interested in price action, I particularly like the rigour with which he describes PA. His objective in the second publication is to define a completely mechanical method to trade any instrument simply based on price action as depicted by a chart. You get a good idea of his thought processes and how the method evolved too.

MC just announced that all (legitimate) users will get free upgrades (a U turn on a U turn ). They are also fairly flexible (or have been in the past) transferring licenses, Having said that I would be a little wary of using betas. Their new releases often break stuff that worked previously and are somewhat buggy (though they are a bit better nowadays than in the early days). I guess it's academic if OEC client will do all you require

This might be of interest http://www.interactivebrokers.com/en/p.php?f=orderTypes# By default IB only simulates an order if the exchange does not support it natively. I think CBoT supports stop and stop limit orders natively.

I have always been a bit suspicious (but everyone thinks I am just a kill joy ). From what you say the client implimentation does not seem to be the issue however I am also a little suspicious of how some of the ninja stuff is written. I do believe that it is neccsaary to use onMarketData to get all the market data to make sure that everything is handled syncronously. Incidentaly Globex alway used to reserve the right to aggregate ticks in fast markets too. I am not sure if that is still the case, the last time I looked on their web site I could not find the bulletin. presumably that would not account for differences between data providers. Be very interested to hear what you find out FT

This is an age old quandry FT. A couple of thoughts If you are going to partial out I would still look at market structure to determine stops targets and RR and see if it is acceptable. So, if placing P1 at the nearest previous swing high and stop loss at the previous swing low, does not give you enough absolute tics profit and more importantly aceptable RR..... pass the trade. If you are partialing out record information for each leg seperately so at a later stage you can analyse things seperately. You might discover for example that you are more comfortable with the initial P1 'scalps' or that you would be more profitable taking 2 off at P1 and 1 off at P2. Connversley you might discover the P1 'scalps' are break even or marginal and that the P2's and P3's make the payrol. By recording all the info seperately and evaluating them as seperate trades with the same entry you will have more idea how things are panning out. Some might find partialling psycologically easier even if strictly by the numbers it is less effective. I guess it's kind of like 'playing with the houses money' I am not sure if that is neccassary a good way of viewing things but if taking a partial helps you run the rest of the position to your ultimate target who is anyone to tell you otherwise Finally P1's will have higher % winners with higher risk lower reward than P2's and P3. Stating the obvious I guess as P1 will always be reached before P2, you will likely discover that lets say 70% of P1's make 1:1 RR and say 30% of P3's make 1:5 RR (made up example). Quite different characteristics.

I don't think so (though I have not tried version 6 yet). You can 'fake it' by plotting HL bars with the plot set to the thickest width. Worth signing up to their forums, people are pretty helpful over there.

I'd second that, I'd even recommend it to people not utilising pitchforks in there trading. Tim's reading of price action is very strong, as his is trade management. Worth a visit for those alone.

Last I looked it actually panned out at about 5 per second. Really the question is would you prefer timely data or complete data? I have not used e$ignal for probably a couple of years now but back then if the market was 'fast' it would get backed up maybe as much as 30 seconds. Of course technology moves on at a pace and things that where an issue then are probably not any more.

Whilst it says get the intraday data with QCollector on a daily basis it is not clear whether that is a limitation of IQ.Feed, QCollector or Gom File converter. My hunch of course is that it is the former which is a shame Anyway thanks again.

Thanks DD for the volatility stuff. RobertM, possibly one of the greatest lessons that newbies might learn with what I call "old fashioned trend following" is that there is no requirement to anticipate what the market is going to do. The assumption is that at some stage markets trend (obviously pick ones that are particularly prone to) all you need to do is make sure you are positioned when they do and rely on sound money management to produce a return. As DD mentioned you need iron cajhones, adequate funding, and a diverse portfolio to trade like this.

hehehe actually its more like does it trade more than 2 in 200ms as that is more like the granularity they aggregate at. Of course if the issue is trading CME products in asia (which does not seem his original intent though he talks about this too) then a different issue again.

Who me? yes of course I would have bought the low print of the day like you ( of course it wasn't a pin bar then but a black candle opening on the high and closing on the low at the time). I would also have bought the low tick of the hammer 3 days ago and sold the high tick yesterday. That's the beauty of hindsight trading of course :D:D So what would have been the basis for your entry (apart from knowing that it had worked out)? Incidentally I used to trade by fading extremes (I mean really rather than 'forum' trades) buying first times at support etc. You can afford pretty tight stops and the RR is generally excellent. Of course there are a whole bunch of other challenges that these early entries present. Of course if your question was addressed to PT99 your guess is as good as mine hence my question. (I thought a clarification might help people get more value out of this thread). As he is trading pin bars and the pin bar did not develop until the 81.20 area I cant see how he could possibly get better than 2.5 to 1 risk reward based on the rather vague entry criteria posted so far.

I'd say re-conciser IB. Do you really need to see every single tick? Very few trading techniques really do. Even if you prefer constant tick charts by selecting a slightly lower value for the constant number of ticks you can get charts that look pretty much the same. Of course if you are doing some sort of advanced order book analytics then you might really need an un aggregated feed. IB data is aggregated rather than snapshot so you get to see every contract traded, so constant volume charts are fine with IB data. Im also slightly confused do you want to trade HK futures from around the world or do you want to trade CME products from Asia? Sounds like both. IB is a great solution anyway having major computer centres in North America, Europe, and Asia. You can set up TWS to connect to which you prefer.

Well to be honest I have never traded a good 'old fashioned' trend following system. I don't think I would have the stomach for it.That is not to say that there is not a lot to be learnt from them. The TT system has some simple features that I found kind neat. 2...% Risk - as we understood it the turtles risked a % of equity per trade/unit. Thats what we did. The volatility adjustment came about from where the stop is set based on an ATR move. It's been a while since I looked but wasn't the number of contracts per unit set based on the underlying volatility of the market in question rather than simple dollar value? Maybe mis remembered that but I could have sworn there was a volatility adjustment somewhere in the recipe (over and above initial stop). Blowfish, can you elucidate us a bit more about the costs of trading using CFDs? what is the average bid-offer, how much do they charge, do they charge for roll-over of cfds, margin costs, etc... I have not 'traded' CFD's for a long long time. Your best bet would be a look at the websites of some of the companies that offer them. They will have all ther pertinent info. One that springs to mind is CMC (though that's not a recommendation!)

Actually there is one other option that is to use uptick/downtick as a proxy for order flow. Whether it is a better or worse proxy than delta and if so by how much is debatable.

Thanks deanz. I am familiar with Gom's excelent work. I guess I need to speak to DTN.IQ to find out what sort of bid ask history they can offer. Thinking about it I would probably require a week rather than a day. I often will have 2 or 3 days off completely impromptu. (Actually just had 5 where I have barely touched the computer) Even so it looks like a rather tedious data management task though possibly scriptable. Does NT 7.0 store historical bid ask data in it's own data base? I really am not that keen on maintaining a separate GOM data base, sure it works well but it is not really robust and scalable. I think if I was going to pursue this seriously (again) I would probably go with neoticker (again) .

Actually looking at it again seems like it is closer to 2.5 or 3 to 1.

Are you OK with a 2:1 Risk Reward? assuming you entered on close and have a stop just beyond the tail of the pin bar?

Hi PT99 I hope you will see that my posts have not been hostile. I would refer what I said around post 25 and one a couple earlier. It is against the rules of the site to keep mentioning a commercial offering however good the material is. You do seem to manage to mention him almost every post. This is one of the reasons you are encountering hostility. Post the charts fine. Keep mentioning xyz service is not. It is probably worth repeating even though Nails stuff is inexpensive it appears to me to be identical to stuff previously presented for gratis by trader dante and James16. (not that anyone has a monopoly on price action) This is a good thing a) because there is no need for a purchase and b) because trader dante showed the efficacy of the method with a real account in 'real time' and even secured a job at a prop shop on the back of it. How you feel about someone packaging public domain material and selling it is down to the individual I guess.

Often times the 'thrust' is what actually breaks out. In this case would you tend to wait for a correction or rest after there has been this price impulse?

Interestingly he went back to the bucket shops a couple of times too