Slippage with different brokers

[Dhalsim]

I have the same strategies running with Tradestation and Multicharts.

The Multicharts is connected to an Interactive Brokers accounts, OEC account and Dorman trading account.

All the brokers used trade the same systems hosted locally in Chicago and this also helps me determine which broker provides best execution and slippage. Tradestation platform seems to be the best and the reason might be because when the signal is generated it is sent to Tradestation order execution very quickly, faster than Multicharts so it is always first in line for limit orders.

My system is built around the fact that either we get a complete fill or nothing. Therefore, should a limit order only partially fill it will convert to a market order within 10 seconds if we still don't get a complete fill. My problem is that as the contract size has increased i am experiencing major slippage. Examples of slippage i am getting:

(All examples based on limit order which executes as market should it not fill within 10 seconds)
- ES trading 200 contracts i average 1 tick slippage on entry, 1 tick on target, 1 tick on stop (this is for a system which trades on the opening minute of the day)
- YM trading first 10 mins of day we get atleast 2 ticks slippage on entry, 2 ticks on tgt, 3 ticks on stops.
- NQ trading systems (trade during all times of day) - 2 ticks on entry, 2 on target, 2 on stops.
- TF Russel - atleast 3 ticks on average for all orders.

Therefore, the slippage is killing me and i have reached scalability for these 1 min chart systems.

- Is there a way around this, i am considering just changing everything to market orders and trying to aim some positive slippage on some occasions?
- Does anyone trading big size on a daytrading setting know what the max scalability would be on some of these markets?
- Do you think changing everything to 2 seconds to convert to market order can improve things?
- Any ideas on better order execution (keep in mind it has to be all in at once or nothing)?
- Also, i have noticed guaranteed stop runs especially during slow periods as we might have 250 contracts resting limit orders which are all show on the book. For those who think other algo's or traders don't target your stops has not traded bigger size.

Any help would be appreciated as the system performance has deteriorated with the big increase in size.

Thanks in advance (also this is a serious thread so no questions about the size being traded, i have a couple of big investors trading with a lot of margin so we sit around the 200 -250 contract range).

 

[nonlinear5]

My system is built around the fact that either we get a complete fill or nothing. Therefore, should a limit order only partially fill it will convert to a market order within 10 seconds if we still don't get a complete fill. My problem is that as the contract size has increased i am experiencing major slippage.

IB recently introduced a new type of order, called "Adaptive Order":
https://www.interactivebrokers.com/en/index.php?f=19091

Presumably, this is to accommodate clients like yourself, who move large sizes, and want to minimize slippage. Have you looked into it yet?

 

[birzos]

That depends on your strategy whether the slippage is material, however let's take it as it is as obviously more details would be needed otherwise. Limit orders have an inherent problem, you're telling the broker more information that you probably want them to have.

As you're trading 1mn all sorts of issues occur, we are able to trigger their algos to delete our orders at these timeframes, it's not clear if you are using auto stop orders. There are technical issues but you are also finding a business problem in the way the markets are structured. We have our own trade management system to route everything around the brokers to avoid these issues.

Market orders will temporarily solve the problem, but you will find other issues become visible in short order. You need to understand that the markets are structured to generate income from other peoples losses and by provider fees. Lots of words, here's the solution.

You want to use retail and mid-market brokers to generate returns in the upper left, but by using those brokers you need to go through the right via the flow of money. What you need to do, and what we have done, is build a bridge (Money Flow) between your brokers and platforms and high income strategies avoiding the right. You have to find your way through the maze, if you manage to build it that bridge will fail from time to time and you need to find workarounds.

You have a financial markets structure problem, not a technical nor strategy problem, don't worry, other people are also starting to have the same problem trying to upsize their positions. The solution is less is more and you need to manage everything via your own systems, not theirs.

How you do that is where the complexity starts, having been an architect for private banks, institutions, and funds I can tell you that you have two options. You can either fix it for today and have a problem tomorrow working from the bottom up, or pause for today and build for tomorrow working from the top down but having to pay for that knowledge and services.

The third is to build for today and tomorrow but you don't have my knowledge, otherwise you would have solved it like we already have. Move to market orders and see how you get on from there working from the bottom up, likely you'll find other 'anomalies' popping up soon after.

 

[tommcginnis]

Total per-day volume is readily available.
Volume per-minute and trades per-minute is also available (if with a little work).
You need to graph those things, along with your slippage, *today*.

Your solution will be readily apparent.

Share this with your clients *after* you are done.

 

[hoppla]

Trading 200 contracts and having 1 tick slippage on average in ES is not bad (but not great too). A former client moved away from Multicharts to a customized software implementation for exactly that reason. The options Multicharts and Tradestation provide are not very slippage friendly. By "customizing" the execution they've been able to cut ES slippage by about 30%. In other markets slippage savings were even a lot better - ie your NQ, YM, and TF slippages are pretty extreme (as in extremely bad). You could save a lot there by moving away from Multicharts/ Tradestation towards a custom software solution.

 

[Gotcha]

I have the same strategies running with Tradestation and Multicharts.

The Multicharts is connected to an Interactive Brokers accounts, OEC account and Dorman trading account.

All the brokers used trade the same systems hosted locally in Chicago and this also helps me determine which broker provides best execution and slippage. Tradestation platform seems to be the best and the reason might be because when the signal is generated it is sent to Tradestation order execution very quickly, faster than Multicharts so it is always first in line for limit orders.

My system is built around the fact that either we get a complete fill or nothing. Therefore, should a limit order only partially fill it will convert to a market order within 10 seconds if we still don't get a complete fill. My problem is that as the contract size has increased i am experiencing major slippage. Examples of slippage i am getting:

(All examples based on limit order which executes as market should it not fill within 10 seconds)
- ES trading 200 contracts i average 1 tick slippage on entry, 1 tick on target, 1 tick on stop (this is for a system which trades on the opening minute of the day)
- YM trading first 10 mins of day we get atleast 2 ticks slippage on entry, 2 ticks on tgt, 3 ticks on stops.
- NQ trading systems (trade during all times of day) - 2 ticks on entry, 2 on target, 2 on stops.
- TF Russel - atleast 3 ticks on average for all orders.

Therefore, the slippage is killing me and i have reached scalability for these 1 min chart systems.

- Is there a way around this, i am considering just changing everything to market orders and trying to aim some positive slippage on some occasions?
- Does anyone trading big size on a daytrading setting know what the max scalability would be on some of these markets?
- Do you think changing everything to 2 seconds to convert to market order can improve things?
- Any ideas on better order execution (keep in mind it has to be all in at once or nothing)?
- Also, i have noticed guaranteed stop runs especially during slow periods as we might have 250 contracts resting limit orders which are all show on the book. For those who think other algo's or traders don't target your stops has not traded bigger size.

Any help would be appreciated as the system performance has deteriorated with the big increase in size.

Thanks in advance (also this is a serious thread so no questions about the size being traded, i have a couple of big investors trading with a lot of margin so we sit around the 200 -250 contract range).

I fine this post funny, even though you said its serious, because you expect the general yahoo ET community to actually have a clue? If you are really trading 200 ES contracts, then hardly anyone here has a clue. If you know who has a clue, meaning you know you is legit, then just ask them. What you will get here is replies from a bunch of wannabes. I'm sure Handle123 will provide a reply since he seems to say that he manually day trades the ES with 400 or 450 contracts (and he apparently can use only a 3 tick stop so he is your man for beating even the algos!!!).

The other thing I don't understand is your 10 second rule. You say you're trading the first minute after the open, and you have a system which will buy or sell after a limit order not being filled for 10 seconds? Jesus, in the opening minute, 10 seconds is a really long time. How on earth can you sit and wait to be filled, and then just randomly go in at the market 10 seconds later when price will likely be nowhere close to your limit order? If price is really sitting on some level and you're not getting filled on a limit order, why not just Buy the Ask, if looking to go long? This should get you filled right away, and you're also limiting your risk of slippage because in case the price has moved past the ask, it should no longer fill since you were only willing to pay the ask as the time and not a full out market order.

Seriously, asking everyone here if you should convert to 2 seconds instead of 10 to hit the market order just sounds like such an amateur question for someone trading 100 contracts. If you were even trading 10 contracts you shouldn't have any questions for the forum community, never mind trading 100 contracts.

 

[hoppla]

@Gotcha you'd be surprised how many old school CTAs and traders with decent size AUM still trade like this and are just coming around to focusing more on their slippage.

 

[Dhalsim]

Total per-day volume is readily available.
Volume per-minute and trades per-minute is also available (if with a little work).
You need to graph those things, along with your slippage, *today*.

Your solution will be readily apparent.

Share this with your clients *after* you are done.

This is something we all understand. However, the system generates a trade during a low volume period but it has a positive expectancy then onbviously i stil

I fine this post funny, even though you said its serious, because you expect the general yahoo ET community to actually have a clue? If you are really trading 200 ES contracts, then hardly anyone here has a clue. If you know who has a clue, meaning you know you is legit, then just ask them. What you will get here is replies from a bunch of wannabes. I'm sure Handle123 will provide a reply since he seems to say that he manually day trades the ES with 400 or 450 contracts (and he apparently can use only a 3 tick stop so he is your man for beating even the algos!!!).

The other thing I don't understand is your 10 second rule. You say you're trading the first minute after the open, and you have a system which will buy or sell after a limit order not being filled for 10 seconds? Jesus, in the opening minute, 10 seconds is a really long time. How on earth can you sit and wait to be filled, and then just randomly go in at the market 10 seconds later when price will likely be nowhere close to your limit order? If price is really sitting on some level and you're not getting filled on a limit order, why not just Buy the Ask, if looking to go long? This should get you filled right away, and you're also limiting your risk of slippage because in case the price has moved past the ask, it should no longer fill since you were only willing to pay the ask as the time and not a full out market order.

Seriously, asking everyone here if you should convert to 2 seconds instead of 10 to hit the market order just sounds like such an amateur question for someone trading 100 contracts. If you were even trading 10 contracts you shouldn't have any questions for the forum community, never mind trading 100 contracts.

Like i mentioned before this is a serious thread - why try to ruin it with your insinuation that just because i am trading size i should know everything about execution when that is not where my skill lies. Read the post it says i have some very large investors and collectively the system is trading well over 200 contracts (close to 300 right now).

This move to bigger size has been a quick transition in the last year where i went from about 50 contracts (with no big slippage issues) to now well over 200. My skill only lies in developing systems with positive expectancy and that is where i currently excel. I have never had to worry about things like slippage on limit orders when designing my systems as we are already trading very liquid futures markets and at most getting 1 tick slippage here and there.

I need to learn many aspects regarding execution or my systems will fail in the long run.

Also, regarding the 10 second rule to hit a market order if limit is not entirely filled, this was back tested in the past when i was trading much smaller size and it showed better performance than hitting a market order straight away after limit is not filled. Also, to clarify the trade is not in the first minute of the day but usually between 5 and 10 minutes. I have many restrictions around execution due to the fact i am using Tradestation and Mulitcharts and many different machines for different investors.

Why does everyone assume that a positive trader excels all aspects of the game? This is a learning game where you move up certain steps. You don't have to be a wizard of execution to have a profitable system but you do need to begin to learn this aspect of trading once you hit a decent size (which is exactly what i am trying to do).

Please lets stop acting like nobody in ET knows anything as this has been a good learning resource for me in the past.

 

[jjw]

This is something we all understand. However, the system generates a trade during a low volume period but it has a positive expectancy then onbviously i stil


Like i mentioned before this is a serious thread - why try to ruin it with your insinuation that just because i am trading size i should know everything about execution when that is not where my skill lies. Read the post it says i have some very large investors and collectively the system is trading well over 200 contracts (close to 300 right now).

This move to bigger size has been a quick transition in the last year where i went from about 50 contracts (with no big slippage issues) to now well over 200. My skill only lies in developing systems with positive expectancy and that is where i currently excel. I have never had to worry about things like slippage on limit orders when designing my systems as we are already trading very liquid futures markets and at most getting 1 tick slippage here and there.

I need to learn many aspects regarding execution or my systems will fail in the long run.

Also, regarding the 10 second rule to hit a market order if limit is not entirely filled, this was back tested in the past when i was trading much smaller size and it showed better performance than hitting a market order straight away after limit is not filled. Also, to clarify the trade is not in the first minute of the day but usually between 5 and 10 minutes. I have many restrictions around execution due to the fact i am using Tradestation and Mulitcharts and many different machines for different investors.

Why does everyone assume that a positive trader excels all aspects of the game? This is a learning game where you move up certain steps. You don't have to be a wizard of execution to have a profitable system but you do need to begin to learn this aspect of trading once you hit a decent size (which is exactly what i am trying to do).

Please lets stop acting like nobody in ET knows anything as this has been a good learning resource for me in the past.

It is my understanding that one of your FCMs has all traffic go through its data center in Connecticut. The latency incurred because of it distance from the CME's data center (which is in Aurora, Il) may be a significant contributing factor to your slippage. If the logic to convert the balance of your orders to market resides in your trading app (this is the case for MC, I do not know about IB) then the time taken by your app and its distance from the system to which it is connected may also be contributing to your slippage. Try using our system and its "convert to market after" feature through R | Trader. We implemented this on our severs which are located at the CME's data center in Aurora.

 

[comagnum]

Try consulting with the tech support at TradeStation - They have a low latency/ high performance network for institutional & HFT -colo'd at the CME and globaly via FCM360.

I use 'fill or kill' which are priority - don't want to be sitting in the LOB (limit order book) - sitting duck for HFTs that prey on these passive orders. Not trading the size you are, although I could intra-day but not overnight but maybe one day I will. So you will have do your own research on that - just an idea.

http://www.fcm360.com/financial-industry-solutions/exchange-connectivity/