Multicharts 8.5 w/LMAX - first week

random12345

Established member
Messages
793
Likes
280
Thought I'd post my experiences of the release variant of 8.5 this week to save other people trouble:

- I didn't experience a single data freeze this week, but it doesn't mean they fixed this problem, I'd still restart every night if money is important to you. However I don't trust this new shutdown strategy resumption feature... it didn't seem right on mine (the strategy re-opened thinking it was flat and zero, when it wasn't). I will still personally use the manual strategy position set when activating auto. MC has always been crap when it comes to advanced auto trading features - I always assume it has it wrong if a shutdown has ever occurred.

- They have changed 1 ct to a micro ct with the LMAX plugin, which is good since occasionally I would get a fill on my exit for most of my position except a few micro lots and MC 8 was unable to cope with this, leaving the pitiful remainder of the position open and messing up the trading report.

- Again, the bloody pt resolution has changed for certain pairs, likely due to the support of micros. EURUSD, EURJPY and USDJPY have all dropped by a decimal point so 70 pts on the EURUSD now = 7 on a MC profit target amount. I failed to alter the EU for my trailing stop this week and I drew down as a result since it was looking for a floor activation 10 times higher than it should have been and since it was a ranging market until Thursday, I dumped plenty of cash. Normally happens in some form or another with an upgrade.

- The trailing stop, for me, is completely borked for optimization. It works fine on live and strategy performance, but not at all on optimization. It gives very strange results. I have informed support as this is a big deal for many of us.

- The feed is a bit smoother with this variant through LMAX, but far from perfect (annoyingly still not as smooth as the prorealtime/IT-Finance skin they have) - so I still advise use of broker side stop orders for position entry. Though I would always advise those...

- Optimization seems a tad faster. Not much though...

- They have improved the commission on backtesting meaning you don't have to derive an average value on the average notional for your data set any more if your broker charges basis pts on the underlying - VERY helpful.

- I'm yet to use all the Delta volume malarky stuff.

- I believe they have improved Renko so that the backtest formation is true to live this time round (it was always a little off), I'm yet to test this though as I don't use Renko any more.
 
Good writeup.
Do you you use EasyLanguage or the MC.NET variant?
I'd guess .net works with LMAX as well?

MC. NET would allow me to easily port stuff from Ninja to MC
as they are both C#.
LMAX is the main attraction of MC for me.
What has your time been like on LMAX?
 
Good writeup.
Do you you use EasyLanguage or the MC.NET variant?
I'd guess .net works with LMAX as well?

MC. NET would allow me to easily port stuff from Ninja to MC
as they are both C#.
LMAX is the main attraction of MC for me.
What has your time been like on LMAX?

I use the EL setup as I was able to get my underlying function to work on EL as I wanted in the end, though I could be tempted to eventually move to .NET depending on how things go this year - I am just using a mid floor, high % trail to exit most of my trades now, which is inelegant. I understand all broker plugins work the same with .NET as with EL. The main advantage of MC for me is the same as you - LMAX. MC isn't a bad package, but I don't think it does a single thing better than any other software out there! Ninja is overall a superior package.

Having said that, I wasn't particularly happy that they borked the optimisation during this upgrade cycle... it had been in beta for a long time and this seemed pretty fundamental. They have now fixed this according to support.

I've enjoyed my time with LMAX - my account manager is very good and withdrawals are fast. They use Worldpay to handle transfers though and therefore their bank tfr deposits tend to take over a day, which is a shame. Their execution, as far as I've noticed, is at least as good as IBs. Stop orders routinely get filled at exactly your price even during news bars. I am very happy with them, it's nice to have a UK based broker with their business model and an API feed. Their CFDs are a bit pricey though, but I don't trade cash indices. Also some of their Australasian pairs seem to have a rather dubious spread given their ECN claim... I'd have to check the NZDEUR again before I gave them a full thumbs up.

Also if you trade over 50 million you get a discount on the 2.5 basis pts default (I was paying 1 basis pt on IB) and this can be individually negotiated. Credit to Shakone for this tip, they didn't mention it to me!
 
I use the EL setup as I was able to get my underlying function to work on EL as I wanted in the end, though I could be tempted to eventually move to .NET depending on how things go this year - I am just using a mid floor, high % trail to exit most of my trades now, which is inelegant. I understand all broker plugins work the same with .NET as with EL. The main advantage of MC for me is the same as you - LMAX. MC isn't a bad package, but I don't think it does a single thing better than any other software out there! Ninja is overall a superior package.

Having said that, I wasn't particularly happy that they borked the optimisation during this upgrade cycle... it had been in beta for a long time and this seemed pretty fundamental. They have now fixed this according to support.

I've enjoyed my time with LMAX - my account manager is very good and withdrawals are fast. They use Worldpay to handle transfers though and therefore their bank tfr deposits tend to take over a day, which is a shame. Their execution, as far as I've noticed, is at least as good as IBs. Stop orders routinely get filled at exactly your price even during news bars. I am very happy with them, it's nice to have a UK based broker with their business model and an API feed. Their CFDs are a bit pricey though, but I don't trade cash indices. Also some of their Australasian pairs seem to have a rather dubious spread given their ECN claim... I'd have to check the NZDEUR again before I gave them a full thumbs up.

Also if you trade over 50 million you get a discount on the 2.5 basis pts default (I was paying 1 basis pt on IB) and this can be individually negotiated. Credit to Shakone for this tip, they didn't mention it to me!

Cheers for the detailed reply, I took a closer look at LMAX the other
day as a spread comparison cropped up on another site, hence my post here.
LMAX do seem to be gaining ground in reducing costs, 0.3 seems to be typical EU spread at peak hours which is pretty good to say the least...
I do find MB good, but the LMAX spreads are hard to beat.

I've never used MC, I'm hoping Ninja add LMAX support in NT8.
They are quite strict about which brokers they support.
Why they don't support LMAX when support is growing is anybodies guess.
Not sure if FIX / API hack can get around it yet.

Its all up in the air atm for me, working on taking advantage of MB's rebate /
liquidity provider options to reduce costs.
Far from sure I can overcome the various obstacles and how the impact on the
performance, not something that can be backtested either which makes things harder.

Although the lower direct costs with LMAX would probably cover most of the
savings of MB's rebate scheme without doing anything.
Costs are far from a game breaker for me anyway, although cost reduction is never a bad thing...
 
Cheers for the detailed reply, I took a closer look at LMAX the other
day as a spread comparison cropped up on another site, hence my post here.
LMAX do seem to be gaining ground in reducing costs, 0.3 seems to be typical EU spread at peak hours which is pretty good to say the least...
I do find MB good, but the LMAX spreads are hard to beat.

I've never used MC, I'm hoping Ninja add LMAX support in NT8.
They are quite strict about which brokers they support.
Why they don't support LMAX when support is growing is anybodies guess.
Not sure if FIX / API hack can get around it yet.

Its all up in the air atm for me, working on taking advantage of MB's rebate /
liquidity provider options to reduce costs.
Far from sure I can overcome the various obstacles and how the impact on the
performance, not something that can be backtested either which makes things harder.

Although the lower direct costs with LMAX would probably cover most of the
savings of MB's rebate scheme without doing anything.
Costs are far from a game breaker for me anyway, although cost reduction is never a bad thing...

Yeah I remember Ninja commenting that they were considering it and it never went further. Shame.

I use the UJ for short term bets - max of 36 trades a day and the LMAX spread and fill is consistently bloody good.

I use long term trading for the EU so the cost doesn't really matter to me as much there, though of course a tighter spread guarantees the performance of static stops and floor activated trails much better... and I currently use both in the EU.
 
Yeah I remember Ninja commenting that they were considering it and it never went further. Shame

Tell me about it.
I found out that you can't even hack the API to connect to
an unsupported broker, as basically NT7 does not have a full API...
Found this on support forum:
connecting "unsupported" Broker - NinjaTrader Support Forum

Only way left, and its a bloody hard long shot, would be to entirely
de-compile the NT source code and add the API functionality,
which I certainly won't be doing, even if it is possible, its beyond me.
 
Tell me about it.
I found out that you can't even hack the API to connect to
an unsupported broker, as basically NT7 does not have a full API...
Found this on support forum:
connecting "unsupported" Broker - NinjaTrader Support Forum

Only way left, and its a bloody hard long shot, would be to entirely
de-compile the NT source code and add the API functionality,
which I certainly won't be doing, even if it is possible, its beyond me.

Have you considered IB or do you only like true tick data?
 
Have you considered IB or do you only like true tick data?
The true tick data issue is not a big deal really, not for what I'm doing anyway.
Currently I'm better off with MB due to the greater flexibility with pos sizing.
http://www.trade2win.com/boards/trading-journals/140032-my-journal-3-a-349.html#post2068566
Costs are similar as well, no partial fill comms traps either.

I'm not trading at any significant size yet, so that greater flexibility
suits me, starting roll of 50k with another 25 to drip feed in on a per trade %
basis.
I'm still relatively new to all this, summer 2011 I thought API was something to do with oil...
I'm happy with MB aside from wanting LMAX costs.
 
The true tick data issue is not a big deal really, not for what I'm doing anyway.
Currently I'm better off with MB due to the greater flexibility with pos sizing.
http://www.trade2win.com/boards/trading-journals/140032-my-journal-3-a-349.html#post2068566
Costs are similar as well, no partial fill comms traps either.

I'm not trading at any significant size yet, so that greater flexibility
suits me, starting roll of 50k with another 25 to drip feed in on a per trade %
basis.
I'm still relatively new to all this, summer 2011 I thought API was something to do with oil...
I'm happy with MB aside from wanting LMAX costs.

Well be careful what you wish for ;) - just as we were singing LMAX's praises, they seem to have technical difficulties right now and have missed their own 17:05 NY 'open' and have even taken the website interface down - I was tempted to heavily short the Euro on Friday night on the basis that at best the Italy news would be poor from the conference, the Cyprus news would have been the icing on the cake. Lucky for me I didn't as I'd be incredibly annoyed if I couldn't see live market data right now given what I read about the pre-market...

I had a feeling this would happen because they took the API portal completely down over part of the weekend and they normally leave that open.
 
Well be careful what you wish for ;) - just as we were singing LMAX's praises, they seem to have technical difficulties right now and have missed their own 17:05 NY 'open' and have even taken the website interface down - I was tempted to heavily short the Euro on Friday night on the basis that at best the Italy news would be poor from the conference, the Cyprus news would have been the icing on the cake. Lucky for me I didn't as I'd be incredibly annoyed if I couldn't see live market data right now given what I read about the pre-market...

I had a feeling this would happen because they took the API portal completely down over part of the weekend and they normally leave that open.
Ouch, then again, most platforms and brokers screw up now and again.
If that type of thing happened every week, that would be a different story.
All the same, don't blame you for being hacked off, hell of a time for it to happen for sure...
 
Ouch, then again, most platforms and brokers screw up now and again.
If that type of thing happened every week, that would be a different story.
All the same, don't blame you for being hacked off, hell of a time for it to happen for sure...

The only real reason I'm annoyed is because it means I'll have to check on it more than I normally would this week - crappy broadband, unreliable software and flaky brokers can really add up as a cost.

I don't think I'll have a position in the EU for a number of days now anyway unless it trades sideways for the whole of Monday... biggest gap down I've seen in a while. My next auto entry point is currently miles away from the current price.
 
The only real reason I'm annoyed is because it means I'll have to check on it more than I normally would this week - crappy broadband, unreliable software and flaky brokers can really add up as a cost.

I don't think I'll have a position in the EU for a number of days now anyway unless it trades sideways for the whole of Monday... biggest gap down I've seen in a while. My next auto entry point is currently miles away from the current price.
Residential broadband?
Leased line may be better:
Leased Lines & IP VPN | Business | Zen Internet

Also maybe a remote server:
Colocation | Data Centre | Hosting | Business | Zen Internet
Dedicated Servers | Dedicated Hosting | Business | Zen Internet
Our Data Centre | Hosting & Domains | Business | Zen Internet
That would be fine as long as its not HFT.

As for EU, manual intervention, or at least disabling any potential long exposure
is what I'll more than likely do.
Was quite a gap, glad I don't do overnights or weekends...
I always close out 20 mins before rollover each day.
 
Residential broadband?
Leased line may be better:
Leased Lines & IP VPN | Business | Zen Internet

Also maybe a remote server:
Colocation | Data Centre | Hosting | Business | Zen Internet
Dedicated Servers | Dedicated Hosting | Business | Zen Internet
Our Data Centre | Hosting & Domains | Business | Zen Internet
That would be fine as long as its not HFT.

As for EU, manual intervention, or at least disabling any potential long exposure
is what I'll more than likely do.
Was quite a gap, glad I don't do overnights or weekends...
I always close out 20 mins before rollover each day.

Tell me about it. I have been trying to get a leased line for some time... it's horsesh*t because I live next to Butler's Wharf and there are technology companies in there...

Might look into a dedicated server, but my technology problems have actually been minimal... the only outage I've had in the past 12 months was due to MC data feed freezing, which is a bug in the software due to a memory leak (now fixed I believe), but as I'm sure you know, it's something you live in fear of with automated trading regardless.

I hold overnight, but not weekends - though I've mentioned in a post before that most momentum or mean based automated traders - channels, moving averages, keltner, supertrend, anything with a TEMA... etc etc etc etc usually coincidentally pick the right direction before Friday closes for the next gap. Over 5 years of data I have an 80% success rate in picking those... but I still don't trade them because it's the most naked of naked trading and if a Lehmans happened over the weekend... yeah.
 
Tell me about it. I have been trying to get a leased line for some time... it's horsesh*t because I live next to Butler's Wharf and there are technology companies in there...

Might look into a dedicated server, but my technology problems have actually been minimal... the only outage I've had in the past 12 months was due to MC data feed freezing, which is a bug in the software due to a memory leak (now fixed I believe), but as I'm sure you know, it's something you live in fear of with automated trading regardless.

Could be they are saving the fibre optic BW capacity for the tech companies...
Pain if thats the case, I don't blame you being reluctant for going the data centre
route.
Its only worth it if you pick a good one with techs that won't tell you
its your fault when one of their 8 year old deathstar drives starts to fail...
Zen are a good outfit, had dealings with them before.

I hold overnight, but not weekends - though I've mentioned in a post before that most momentum or mean based automated traders - channels, moving averages, keltner, supertrend, anything with a TEMA... etc etc etc etc usually coincidentally pick the right direction before Friday closes for the next gap. Over 5 years of data I have an 80% success rate in picking those... but I still don't trade them because it's the most naked of naked trading and if a Lehmans happened over the weekend... yeah.
Thats my primary motivation for closing at weekend as well,
rollover close is more about sleeping easy and no rollover costs are an added bonus.

Never used any of them, or had much success anyway.
Been messing about with linear regression for last week,
not with much success yet, has to be said.
Nature of the beast though.
 
Never used any of them, or had much success anyway.
Been messing about with linear regression for last week,
not with much success yet, has to be said.
Nature of the beast though.

I remember reading a study of LR, I think it was a journal about the end point moving average. I remember thinking it would have institutional use, but limited retail application - zero lag stuff like that would make anyone rich in about a month.

That is, if there was no such thing as commission, spreads and slippage. :whistling
 
I remember reading a study of LR, I think it was a journal about the end point moving average. I remember thinking it would have institutional use, but limited retail application - zero lag stuff like that would make anyone rich in about a month.

That is, if there was no such thing as commission, spreads and slippage. :whistling
Pretty much sums up what I was starting to think myself :cheesy:

Plenty of other ideas, next one will hopefully be ES based on tick volume.
Other than that, adapt what I have for other pairs with a low correlation,
whilst attempting to avoid doubling up on USD exposure.

May even look into a long only AUDJPY carry type method.
Although the unwind risk can be savage atm.
All still just ideas, whether they have any merit when automated
remains to be seen.
Anyway, good discussion, time for me to disappear :)
 
Top