MC6 Beta 1 Feedback

I was trying to float the tool bars

To move any toolbar outside the main MC window, you need press and hold Ctrl before you start moving it. Toolbar locations are recorded and they will appear in the same location next time you start MultiCharts.

HERE IS THE PROBLEM with GLOBAL MODE:
When I do not click “Global Mode” the replay correctly executes the “As is (10 Seconds)”. However, if I do click “Global Mode” the replay incorrectly switches to “As is (1 minute)” and I can not get it back to “As Is (10 Seconds)” which is what I will always want.

The Global Mode was designed for a simultaneous and synchronized data replay on all open charts and scanner windows within the running instance of MultiCharts.

In Global Mode, there are only two types of playback: “Minute by Minute” and “Day by Day”. The reason for this is that many open charts may have different resolutions, and it’s not clear for the program which one should be used as the Playback Resolution.

The same logic applies to Playback Resolution in the scanner, and charts with several data series.

As soon as the Global Mode button is pressed, JumpToBar is executed on all charts, specifying the same time of day, and taking into account Time Zone differences.

The 10 second resolution will run fine on one chart only, as long as you don’t click the Global Mode button. As soon as you do, it will jump to “Minute by Minute” and “Day by Day”. As a side note, it possible to specify more than one minute by going to the StepForward or StepBackward menus and choosing “10 Minutes Forward” or “10 Minutes Backward”

Can you use the replay while collecting RT data, or only in off line mode?

You can use Playback and Collect Data simultaneously. The Collect Data feature is not tied to Playback in any way. As a side note, there is a limitation that if a chart has Playback turned on, you cannot enable Autotrading on that chart.

Anyone else have a problem with Renko charts not going back far enough in history?

You can specify any request for a Renko chart, but the chart will only display 1,000,000 Renko boxes of historical data at maximum. As real-time data arrives, more boxes will be added.

The Renko chart is constructed only based on ticks. So, if there are no ticks in the database, then the amount of historical data will be limited by the number of historical ticks delivered by the data feed.

Also, for non-regular chart types, there is no backfilling, so it takes some time to construct a data series.

I was forced to reinstall my PC (I bought a new graphic card that, apparently, my motherboard didn't approved, making windows, sound card, internet connection going nuts. Don't you hate when you need to buy a complete new system because the old one don't work with the latest version of some type of hardware and you cannot find previous versions anywhere?), and I noticed one thing.

Every night my backup software saves the Database and the StudyServer folders into a usb flash drive.
Today, after installing MC 6 I copy/paste the folders and went to PLE to compile the studies and see if everything was OK.

First, why aren't all the folders that we created being saved?
It's quite painful for people that have hundreds of indicators and all of them separated by folders to have to create all again and separate all indicators/functions/studies, again, into each of them.
The fact that there's no cut/paste option, it's even more painful to select the indicators that are on the top of the list and drag them onto the bottom, where the created folders are.

Another thing that would be nice, would be the possibility to save the toolbar customization into a file that we also could save/backup.

Hi Andrew,
Regarding

The 10 second resolution will run fine on one chart only, as long as you don’t click the Global Mode button. As soon as you do, it will jump to “Minute by Minute” and “Day by Day”.

As I basically said before, The global mode (if it forces the "Speed Control Chart" to a higher resolution than the lowest resolution chart) it esentially renders trigger practice impossible (probably for most and more likely all discretionary day traders at least) since they create the lowest resolution chart just for that purpose (to get a finer signal in the waves to trigger a good entry price). It might be okay for EOD traders however if their lowest resolution chart is in fact the 1-minute bars but I am going to go back to day traders who are almost always going to have a chart below the 1 minute bars as their lowest resolution chart . During the reality of the trading day (which obviously is what replay is for) the 10 second bars do not shoot out in sets of 6 bars. At best (for trigger practice) the fastest a discretionary trader is going to want to see the 10 second bars move would be 1 bar a second (and even that is 10 times faster than reality to give their skills a good work out). Shooting them out in sets of 6 is a long ways away from reality. I was finding that I wanted to slow it down to one 10-second bar every 5 seconds rather than one 10-second bar every second. Some traders will be using 5 second bars or maybe even tick charts. The same principle will apply here.

The idea of having the "Speed Control Chart" set to a higher resolution than the lowest resolution chart of the replay is probably okay for setup practice (maybe even a great idea if you can do it and it appears that you already have done it). Essentially the trader is analysing the larger picture and trying to project where prices may end up both for entry and for exit. In that case they are not yet trying to analyse the lower levels of waves to try and get a good entry price (triggering might be defined as trying to get a good price without missing the trade). If it is easier (from a programming perspective) to only have one "Speed Control Chart" it might be better to always have it the lowest resolution chart and simply provide much higher speed capability for setup practice. Having said that I am not sure how this would work with tick charts. Tick charts are special in that their idea default normal real life speed on the slider scale would be the speed at which they arrived which hopefully is very close to the same same speed that the trade took place on the exchange (maybe to do an exact relay here the arrival time needs to be used). Maybe discretionary traders who use tick charts do not need replay. For me (after 8 years of very intense almost full time 16/7 study of the market), time is very important (especially at wave tips) which is why I use 10 second bars. I use a wave comparison technique where I put a line from the start of a wave out to the end and clone it to the start of the next wave so I can see how the next wave is progressing on the forward movement relative to the prior wave. It is automatically programmed in EL code to do this based on where I place the arrows at the tops and bottoms of the waves. The line changes color at 75% and 100% and I get a popup notification. I get this on two levels of waves. This is very time based analysis and that is basically why I need time based 10 second bars. I am very happy with my strategy. I will never use a tick chart. Oops! Never say never

Here are a few things to add to the above post that I created offsite.

Regarding the speed control chart, your current feature of changing it when the trader clicks on a chart is a great feature and it makes it questionable as to whether you need to store what the trader last used since it is so easy for the trader to select it each time.

Here is a better wording for something I mention above. Again for trigger practice I figure all traders when doing trigger practice will normally be using the lowest resolution chart for the “Speed control chart” and my current thinking is that for setup practice (which is a bigger analysis of the market done before they get down to the triggering level) they could simply select the lowest resolution for their “Speed Control Chart” and run it at higher speeds rather than select a larger “Speed control chart” to get the greater speed. On the other hand, If I can select a larger “Speed control chart” for setup I will probably ON OCASSION do that rather than select a smaller “Speed Control Chart” at a higher speed. What I am doing is being realistic here in that I realize (at this point in time at least) that it is probably not that easy to program the syncing up of the charts and the best compromise here is to always keep the smallest resolution as the “Speed Control Chart” and allow the user to use much higher speeds on it for their setup practice.

“Speed Slider”
I think it is pretty obvious that the best default for the speed slider after a “Speed control Chart” selection has been made is the normal real life speed of that selection. Examples: Maybe the program could remember the trader's last adjustment. As I mentioned above I found I wanted to run 10 second bars at 1 bar every 5 seconds (double the real speed) rather than 1 bar every second (10 times the real speed). However what I found I was doing is I would run faster then as I started to see the trigger approaching I would bring the speed down.

“Speed Slider” continued - This is a very interesting area:
For tick charts it is the speed that the ticks were coming in I guess that would be your default speed. I have no idea how you would deal with this one (maybe apply a tick arrival time stamp makes sense). I don’t know how you have programmed the replay at this point but it appears that in order to get the bars to look exactly as they did when they came in you have recorded the close values of (open, high, low and close) and stepped based upon these values. To me this seems like a very smart way to go and for time based traders it is perfect.

I am very tempted to outline my top down discretionary technique in detail to make it really clear that (for a day trader who is doing trigger practice) the Speed Control Chart needs to be lower than 1 minute bars and they might want to do this practice in all charts global sync mode too (I found that I was using it last night because I was getting all the GV updates from the larger charts as well and I was able to do a true simulation this way - what you have done is really cool and your sales for sure will go up because of it). However I have already written enough and won't get into the details of my system especially since a lot of traders may start using this thread for a topic other than MC6.0 feed back. On the other hand it may take your sales even higher since traders who can't program well may find MC more useful if they have a good system to practice using the replay (maybe you need a thread on using the replay effectively or practicing different trading systems using the replay).

For the replay it would be useful if the studies could get the replay jump date and time via GVs reserved by MC. I am thinking along the line of info like this.

MC_ReplayStartDate
MC_ReplayStartTime
MC_ReplayStartTime_s

The user could test against the previous values in their studies to detect that a new replay had been started. If these values were zero then the study would know that this is not a replay run or at least that the real time button had been clicked. Maybe zero could be live and real time clicked while MC is off line would send out the last bar on the chart values to these GVs.

Here is what I think I am going to end up doing. The first study on the 10 second bar chart will pick up on the GV that tells it replay mode is in effect. It is set to value 1 in my database program. On the first bar on chart it will set a switch. On the last bar on the chart the above fields will be sent out but also the first time switch will set some GV fields needed by the second study on the chart. The second study will use the above fields and the initialization will have taken place and it will process correctly. It may also have to use the replay mode GV.

Maybe the replay tool bar could have an options button for the above. Maybe reserve words are better. Maybe a replay status (play/pause) would be handy. The advantage of GVs is external programs can detect the replay is in pause mode even if it is stopped where as the study can not ship a reserve word value out to a GV if the study is stopped. All somewhat low prioirty ideas I would think but there you have them for whatever they are worth.

This is sort of interesting. I don't think it was intended but I will mention it.

I was running the replay on 10 second bars. I had the full 1 minute bars showing. I thought I would try (while the 10-second replay was still going) clicking jump and clicking the 1 minute bars in the future of where the 10 second bars were currently replaying then clicking play. What happened was both the 1 minute bars and the 10 second bars were replaying (1 bar per second) but from two different locations. Once the minute bars came to the end of the day it stopped as one would expect and the 10 second bars just kept chugging away as happy as can be.

Now what would be really cool would be if I was smart enough to trade that

This is sort of interesting. I don't think it was intended but I will mention it.

I was running the replay on 10 second bars. I had the full 1 minute bars showing. I thought I would try (while the 10-second replay was still going) clicking jump and clicking the 1 minute bars in the future of where the 10 second bars were currently replaying then clicking play. What happened was both the 1 minute bars and the 10 second bars were replaying (1 bar per second) but from two different locations. Once the minute bars came to the end of the day it stopped as one would expect and the 10 second bars just kept chugging away as happy as can be.

Now what would be really cool would be if I was smart enough to trade that

I tried to "Jump" to tomorrow's chart,
but MultiCharts just stuck at the last bar.

TJ- LOL- that's the Holy Grail, we need Andrew to put that on his "to do" list. I'd pay extra for that feature.

I tried to "Jump" to tomorrow's chart,
but MultiCharts just stuck at the last bar.

I agree, Not being able to see the future has reduced my income too. You can buy crystal balls cheap I think

Here is an interesting issue. I recently changed some of my EL trigger related code to go after the bid or ask rather than the close price, but during replay it isn't available. So I guess a GV needs to be set to switch it back to the close price during replay. I also had to change it to trigger on the last bar of the chart rather than mid bar. For me it is not a big deal. I manually decide when to switch on the triggering on a final bar to bar stall. I am not practicing the specifics of how the bar to bar stall works. Just the manual part.

Lots of interesting issues come up with this feature.

Good job in general on the playback feature.

Please note that there could be significant lag (or misalignment) in the playback given a large array indicator, two instruments (one 12 range-26days back, one 5 min-55 days back) chart that is not apparently CPU dependent. The large array indicator is based on the second hidden 5 min instrument.

The attached animation showed such a problem with @ES, showing an offset display between the top panel and lower indicator panel if the playback was performed using 'Jump' to 16 day prior to current date. The playback does not show such an offset alignment if played from, say, 8 days prior.

Here is a little more detail on something I mentioned before.

I am not really using global mode that much at the moment because I can not step the 10 second bars 1 bar at a time when in global mode. However I have been using it to set up some GV settings which can only be set up with global mode running. I run global mode a few bars then pause and switch back to 10 seconds bars to get the detail I need for trigger practice. When I do this and I click the global button all the other 9 charts line up (it takes maybe 3 to 5 seconds to do this which is pretty good) then there is a further delay before the play button becomes available. Sometimes this further delay is fairly long (30 seconds maybe). Like I said, I mentioned this earlier but I had not noticed that the longer delay occurs after all the charts have been lined up. This seems fairly consistent but sometimes it seems a bit longer than at other times. If it was this way in TSS testing then I guess it is normal. Maybe it is the fact that I have 9 charts in the workspace that leads to this delay. If it is normal maybe the message could change after the actual actual chart line-up.

What I am about to suggest would be helpfull. I often wonder if the global button has created a lock out situation and I start clicking things to try and correct it or get the play button to finally become available.

I noticed that every control on the playback tool has control tip text which the user can see if they put the tip of the cursor just at the very bottom of the control (if they put the cursor in the center of the control they do not see it – this is different than what I am use to since control tip text I am use to uses a “first time show” method). You might want to mention that in your replay manual so they can know about it for the next suggestion. Also for the next suggestion you might want to use the format I use for control tip text which is,

Optionally initials on the control itself if you have room for it.
Control tip text starts with the words for those initials.
After the words there is a dash then the control tip text explains what those words mean and maybe explains the function of the control as part of this. Sometimes I put these explanations in round brackets.

Okay my idea is that you could make room for a small activity box on the playback tool itself to show what the replay is doing when it is busy and the buttons are locked out for a significant length of time (personally I would avoid putting it on the status bar since many users such as myself will not have the activity bar showing so they get more chart space). Room is limited so only have a small activity box. In the box all the user sees is the activity number and the initials and they will change as the activity changes. If they place the cursor at the bottom of this control they can see the control tip text changing which explains what these codes mean and you use the format above I just described. After one time through you will have convinced them that they have to wait and they should be patient. After a while the user will learn to tie the number to a specific amount of wait time, which saves you having to figure out how to estimate it in seconds. Here is a dummy example format. If an activity flashes buy fast they won’t have time to read it so I guess you will have to summarize them properly so they get a general idea.

1-ABC
2-DEFG
3-Whatever.

Maybe total seconds taken so far could be included so they get to know a typical total wait time.

What I am about to suggest would be helpfull. I often wonder if the global button has created a lock out situation and I start clicking things to try and correct it or get the play button to finally become available.

I noticed that every control on the playback tool has control tip text which the user can see if they put the tip of the cursor just at the very bottom of the control (if they put the cursor in the center of the control they do not see it – this is different than what I am use to since control tip text I am use to uses a “first time show” method). You might want to mention that in your replay manual so they can know about it for the next suggestion. Also for the next suggestion you might want to use the format I use for control tip text which is,

Optionally initials on the control itself if you have room for it.
Control tip text starts with the words for those initials.
After the words there is a dash then the control tip text explains what those words mean and maybe explains the function of the control as part of this. Sometimes I put these explanations in round brackets.

Okay my idea is that you could make room for a small activity box on the playback tool itself to show what the replay is doing when it is busy and the buttons are locked out for a significant length of time (personally I would avoid putting it on the status bar since many users such as myself will not have the activity bar showing so they get more chart space). Room is limited so only have a small activity box. In the box all the user sees is the activity number and the initials and they will change as the activity changes. If they place the cursor at the bottom of this control they can see the control tip text changing which explains what these codes mean and you use the format above I just described. After one time through you will have convinced them that they have to wait and they should be patient. After a while the user will learn to tie the number to a specific amount of wait time, which saves you having to figure out how to estimate it in seconds. Here is a dummy example format. If an activity flashes buy fast they won’t have time to read it so I guess you will have to summarize them properly so they get a general idea.

1-ABC
2-DEFG
3-Whatever.

Maybe total seconds taken so far could be included so they get to know a typical total wait time.

If the play button is locked up for a while, a beep once it is available would allow the trader to do somethine else while they wait and get right to the replay once it is ready.

I just completed a full run through a trading day where I was actually trying to practice.

The really nice thing about this replay feature is the ability to leave it running on the lowest resolution chart and with just two button clicks start it again from a new location. Its incredible TSS. I find it best to work backward during the trading day to practice the triggering. It is such a fast feature that if the restart location is out of sight to the left you just do it a few times until you get it within sight then you can zero in on it. Running without global mode is okay because the 1 minute bars and larger bars can be used to view the next location to practice but if working backwards global mode would be okay too.

It even does a recalculate on the studies when the restart occurs which is great. I have all the specialized programming in which detects that Replay is in effect.

I find that with the 10 second bars normally I run it at 1 bar a second but I run it faster or slower for special situations but not for all that long (during practice 10 times faster this is a good standard to try and meet). What would be really useful would be a favorite speed button as well as the + and - buttons. Maybe a right click on that button could be used to set the favorite speed to match the current slider speed and the left click to jump the slider back to the favorite speed.

If all this could be done in global mode just as easily and with the SpeedControlChart being the lowest resolution that would be really cool.

now all we need is a sim engine to go with the playback,
that would be the icing on the cake.

I have been resisting asking this question but if I don't it is going to drive me crazy all day

What is a "sim engine"?

a simulated order processor...

for discretionary traders,
he can click buy/sell during the playback...
as if he is trading in real time.

MultiCharts will then tally up your PnL at the end of the day.

a simulated order processor...

for discretionary traders,
he can click buy/sell during the playback...
as if he is trading in real time.

MultiCharts will then tally up your PnL at the end of the day.

That would be something of this world

With a simulation engine like that, we could almost simulate at a 95% rate or more, the phsycological forces that a discrecionary daytrade is taking every single minute during a trading day.

I would love to have access to that type of "engine" in MC

Regarding your "SIM Engine" TJ,

if I was to run this replay forward in global mode starting back a long ways (or starting back at a database entry if TSS puts that feature in) so I did not know the future (or a student who did not see me do this had no Idea of the future) and just let it run forward at normal speed I could use the database program to create a SIM engine and record the profit and loss of the decisions for the day. If I was to teach and create a test I would probably do that but this would not be possible unless TSS changes the replay to have the default "SpeedControlChart" in global mode as the lowest resolution chart (in my case 10 second bars). I have some of it in place already in that I have a database set up for profit and loss. It just needs some tweaks. Currently I have a system set up where I can show a graphical history of certain decisions made during that period (graphical in that they appear as codes right on the chart itself). It uses a matching record process to bring in the file that has this data. It uses the EL collection MapNN.Next command to match against the bars. I use it off and on. If I had students on a network I would monitor their decisions live.

I won't say much here but The trading system is a top down system where you trade in the direction of a projection that is based off of at least two levels of waves (from the trade wave and up) and triggering is basically done by working down the wave levels (starting from your trade wave level) looking for a break or stall at one level before dropping down to the next level to look for the same thing again. Once these break or stall signals get you down to this very same signal at the tiny wave level you initiate an auto fire trigger to detect a stall at the B Wave level (meaning a bar to bar wave stall as found in the 10 second bars). Repeat bounces are factored in as well. This system has a very good win/loss ratio (losses are typically break even scraches or even 1 unit profit scratches). You can only trade this manually (it is impossible to program because sometimes the number of wave levels of projection changes as does the number of wave levels of trigger - not to mention the bounces as well). You can only program MC's EL to assist you with this system (I have worked fairly hard at trying to figure out how it can be automated and it can't). Therefore the only way you can actually test this system is to test a trader who is trading it or test it manually if you are careful. I am pretty happy with it. It is 99.9% done I figure. Rare changes take about 3 minutes (no EL code programming needed here). I simply change my checklist and it is done. My eyes do what the EL code would have a great deal of trouble doing.

If TSS makes the "SpeedControlChart" change I suggested I may try teaching. We will see. I would like to set up a team trading approach as an alternative if I can find a great student who can trade this system well. I have no doubt that many can do it better than I. The advantage of trading this system as a team is that if one person misses something visually the other hopefully will spot the error.

I have a "visual" system as well- what I do, is trigger an email and audable alert, when a trade is "close" , that gives me time to see if a good signal is setting up.
That way a "student" shouldn't miss a trade, if he knows what to look for.
I also love the new "playback" feature and used it to look at a couple of weeks of ES, over the long weekend . I was very pleased to see that it works with multiple data streams, as long as you're in "global" mode.
It's going to be a big help to visual system traders. I think it can actually be used to "tune" indicators, so they will look the way you want, just prior to a trade.
Great job TSS support !

Hi Steven,

Is your lowest resolution chart at least 1 minute bars (or greater) and do your final trigger signals come from this chart (or at least primarily from this chart). If you are not using a resolution below 1 minute then global mode is fine for you.

John.

Hi Steven,

Is your lowest resolution chart at least 1 minute bars (or greater) and do your final trigger signals come from this chart (or at least primarily from this chart). If you are not using a resolution below 1 minute then global mode is fine for you.

John.

**John: I was trying it out with 5 min and 30 min bars together.

Regarding this problem I reported on the first page,

However I noticed in the replay that the "1 update every 2 secs, 3seconds and 4seconds" is missing. It jumps directly from 1 update every 1 second to one update every 5 seconds yet there are snap positions for these other speeds.

I layed out the exact slider control speeds I am getting when the "SpeedControlChart" as I call it is set to 10 second bars.

1 update every 10 seconds
1 update every 9 seconds
1 update every 8 seconds
1 update every 7 seconds
1 update every 5 seconds
1 update every 1 second (should be 4 seconds)
1 update every 1 second (should be 3 seconds)
1 update every 1 second (should be 2 seconds)
1 update every 1 second
1 update every 1 second (duplicated the one above it)
2 updates every 1 second
2 updates every 1 second
3 updates every 1 second
5 updates every 1 second
10 updates every 1 second
11 updates every 1 second
10 updates every 1 second
11 updates every 1 second
12 updates every 1 second
14 updates every 1 second
16 updates every 1 second
20 updates every 1 second
25 updates every 1 second
33 updates every 1 second
50 updates every 1 second
100 updates every 1 second


The very quick jump up to very high speeds is really good for moving quickly to another practice point in the future. I have started to do this by sliding the contol right to the far right, watching the prices move quickly buy then pulling the slider very quickly back to the far left to stop it. At that point I use the jump button to move back to the exact start point and then if it existed I would use the button for a traders favorite speed.

As far as the useful speeds that are missing shown again below along with the "1 update every 1 second" which is also a good speed to work with,

1 update every 1 second (should be 4 seconds)
1 update every 1 second (should be 3 seconds)
1 update every 1 second (should be 2 seconds)
1 update every 1 second

I am finding 1 update every 1 second at times is just a touch too fast and if a favorite speed button was put on I might set it at "1 update every 1 second" then click the "-" button to back it off to one of the slightly slower useful speeds above. For a difficult trade I might back it off 2 speeds and for easier trades maybe only back it off one speed.

The interesting gaps in the higher speeds makes me curious how these speeds came about. If they were just randomly chosen then maybe some traders would want to fine tune the gaps and request a way to do that. For me including the missing ones is probably all I would ever request.

Hi.

Apparently MC 6.0 is not handling data as MC 5.5.

I used the same database in both MC versions. I copied TSSTORAGE.GDB into my laptop and compared the ES charts.

Please see images attached.
In them I focused on a especific day (26-11-2009, an holiday), but this type of issue is happening all over the database (from 11-12-2007 til today).

Basically, MC 6.0 is not plotting November 26. It jumps from 23:59, November 25 to 12:04, November 27.

As we can see in the images presented, MC 5.5 is plotting November 26.

How can the same database plot different results?
Why MC 6.0 not plotting the holidays? At least November 26 is an holiday. All the other days where this is happening, I didn't confirm if all are holidays.

Hi.

Apparently MC 6.0 is not handling data as MC 5.5.
....

That's a continuous contract.

Are you using the same rollover rules for both charts?

That's a continuous contract.

Are you using the same rollover rules for both charts?

Yes.

I'm using the continuos IQFeed ticker (@ES#).

QM setting are also equal in both MC version.

I check this year, and this is always happening on holidays.
Is there a on/off switch to this?

I can't find it.

Please don't tell me that I need to delete the dates from the Holidays menu in QM. It would be quite awkward.

Anyone else have a problem with Renko charts not going back far enough in history?

Yes.

I am using esignal. For PF, Renko and Kagi charts it seems to get only 14 days of data.

LB looks fine.