IB Trades Executing at Bid and Ask Rather Than Last Price

[carltonp 10]

Hello Traders,

 

I'm using IB as my broker. For some reason, my trades were continually executed at bid/ask rather than last price. For example, I submitted a long trade to be executed at 11824. The trade was executed at 11830 at ask price, even though the last price never reached 11830.

 

Can someone please enlighten me.

 

Cheers

 

Carlton

[Kiwi 16]

You might want to enlighten us first:

- what are you trading?

- what type of orders are you using?

- are you entering them directly on tws and if not do you know what was on tws?

and

- when you say the last price never reached x, the ask and your execution, what do you mean (what data are you looking at?)

[mike21 10]

Carlton,

 

If you put a limit order at 11824, you would have been filled at 11824 or not at all.

 

You obviously used a market order. A market order will be filled at the best bid/ask at the time of the order.

 

Once your order was filled at 11830, that became the 'last print' so 11830 had to have been reached.

[carltonp 10]

  Kiwi said:

You might want to enlighten us first:

- what are you trading?

- what type of orders are you using?

- are you entering them directly on tws and if not do you know what was on tws?

and

- when you say the last price never reached x, the ask and your execution, what do you mean (what data are you looking at?)

 

Gents, thanks for responding.

 

I'm trading mini dow(YM).

 

I'm placing buy-stop orders.

 

I'm entering them directly on TWS.

 

So, yesterday I placed a buy-stop order at 11824. The trade was executed at 11830. I don't have a problem with that. The problem is, the last price never reached 11830, however the ask price did. I've attached a snapshot of the exact trade as it happened yesterday. You will see that I have placed a buy stop at 11824. You will also see that the trade has executed at 11830 at the ask. However, you notice that the last price is 11826. What you won't we is that the actual last price never reached 11830. Normally, my the trades will be executed at the last price and not at bid / ask.

 

I hope this visual helps you to help me understand.

 

Thanks

[Kiwi 16]

A stop order (not stop limit) will execute at the extreme of current range.

 

IB's data feed (not the 5 second bar version but the real time one) shows a price every 100ms. Its the last price in that 100ms and if the extreme is greater it doesn't show. So you can get filled and the price not show up. Not every price is necessarily shown.

 

If you ask IB to confirm it they can check but it will have filled at the extreme.

[SIUYA 11]

another thing to check on TWS.

If submitting some orders and they are filled, the reported price they are filled at can easily be confused with the limit price sent

 

In this example below, I sent a LMT order to buy at 1.302...the actual trades were done at 1.3018 in two tranches.

 

- BOT 50K EUR EUR.USD Cash 1.302 USD IDEALPRO 15:27:29

BOT 20K EUR EUR.USD Cash 1.30180 USD IDEALPRO 15:27:29

BOT 30K EUR EUR.USD Cash 1.30180 USD IDEALPRO 15:27:29

[steveh2009 10]

A buy stop limit order doesn't ensure execution, but it will ensure that you get the price you want or better.

 

What I do with my own trading interface to IB when trading the CL is, if I want in the market immediately but don't want a market order which could slip me 2-4 ticks, I go 1 tick inside the bid on a buy limit or 1 tick inside the ask on a sell limit.

 

If I am staggering buy orders ahead of where the last is in an uptrend (e.g. 20 ticks between entries), I'll put in stop limit orders when my averaged up buy-ins are closer to my initial position abd have the ones farther away with stop order triggers where the slippage isn't going to threaten my overall cost average within the trade.

 

I'll give you an example of an actual pleasant experience I have had with limit orders. I'll have my software set to trigger a limit buy (or sell if going short) on touching, say, a 20 ema on a 5 min chart. The CL can get pretty jumpy on such an event so by the time my limit buy at, for example, hits 95.50 and got executed, I found that I got filled at 95.45...5 ticks better than what I wanted and then price is suddenly at 95.55 or better, 10 ticks to the positive in the blink of an eye. Had I placed that limit order directly in the market beforehand, I would have gotten filled at 95.50.

 

You can see what's going on here. The software sees a last of 95.50, sends the limit order to the exchange. Meanwhile, the exchange sees my limit order at 95.50 while its ask is touching 95.45. I get a much better fill in this case. It helps to keep me from getting stopped out on a sudden surge into support / resistance because my stop loss order, along with a target offset from the actual limit fill price reported by the exchange, is submitted as an OCO order on the execution status info from the entry. It doesn't happen a lot, but when it does, it's usually a good thing (for a change!)

 

Over the long term, if you're executing market orders based on something like X seconds before a bar close, you'll see price favoring you many times too relative to the actual closing price of that bar. You've got a lot of programs out there waiting for the exact close of a price bar (5 min being a common one) before they react so a market order like 3 secs before the close averages out pretty well to that close. I prefer this approach to finding out the closing price first and then entering a limit order because A) you can and will get skipped over at your ideal entry point and B) you're NOT going to like your at-the-market fills if your analysis is right and that closing price was what you wanted to have all along.

Edited December 15, 2011 by steveh2009

[carltonp 10]

Gents,

 

Thanks for responding. I'm putting a call through to IB now, will let you know what they have to say.

 

Cheers

[carltonp 10]

SteveH,

 

That is a fantastic entry method - brilliant. I'm going to try and emulate it and see if it works for me.

 

Excellent

 

Cheers

[mike21 10]

  carltonp said:

Gents, thanks for responding.

 

I'm trading mini dow(YM).

 

I'm placing buy-stop orders.

 

I'm entering them directly on TWS.

 

So, yesterday I placed a buy-stop order at 11824. The trade was executed at 11830. I don't have a problem with that. The problem is, the last price never reached 11830, however the ask price did. I've attached a snapshot of the exact trade as it happened yesterday. You will see that I have placed a buy stop at 11824. You will also see that the trade has executed at 11830 at the ask. However, you notice that the last price is 11826. What you won't we is that the actual last price never reached 11830. Normally, my the trades will be executed at the last price and not at bid / ask.

 

I hope this visual helps you to help me understand.

 

Thanks

 

Carlton,

 

I trade live everyday with IB, and 95% of my entries are with a buy-stop or sell-stop. So maybe I can clear a few things up.

 

First of all, I see you're trading sim, so its possible you got a goofy fill that would not have been filled in a live account.

 

A typical buy stop is a market order placed above the current bid/ask. When the price you set your order at is printed (the last price) your order is sent to the exchange as a market order and your order is filled at the ask. If the market is moving fast, there's potential for slippage.

 

I don't trade the YM but I see it has decent liquidity, so the 6 tick slippage you received is probably unusual.

 

~Mike

[carltonp 10]

  mike21 said:

Carlton,

 

I trade live everyday with IB, and 95% of my entries are with a buy-stop or sell-stop. So maybe I can clear a few things up.

 

First of all, I see you're trading sim, so its possible you got a goofy fill that would not have been filled in a live account.

 

A typical buy stop is a market order placed above the current bid/ask. When the price you set your order at is printed (the last price) your order is sent to the exchange as a market order and your order is filled at the ask. If the market is moving fast, there's potential for slippage.

 

I don't trade the YM but I see it has decent liquidity, so the 6 tick slippage you received is probably unusual.

 

~Mike

 

Hey Mike, do you work for IB? (just joking). When I called their support team they virtually said what you mentioned. However, you're explanation made it really clear for my simple brain to comprehend.

 

As you use IB and TWS, can you tell me if its possible to place a 'buy stop limit order' as suggested by SteveH?

 

Cheers

[SIUYA 11]

  carltonp said:

 

As you use IB and TWS, can you tell me if its possible to place a 'buy stop limit order' as suggested by SteveH?

 

Cheers

 

Try this....its well worth learning the system you are going to use......you might find something new and exciting.

 

http://www.interactivebrokers.com/en/p.php?f=orderTypes

[mike21 10]

No, I don't work for IB, but I've used them for several years.

 

I read steve's post but I'm not sure I completely understand his approach.

 

Typically, a limit order (for a long trade) is placed below the current price and is filled if price moves down to your limit price. But just because price prints at your limit doesn't mean your order will be filled. Limit orders are time-stamped and filled in the order they're received at the exchange. (the queue) So when price prints at your limit -- first, market orders are filled, then the limit orders are filled in order. If there are not enough sellers, price will move back up without filling all the limit orders.

 

Personally I don't like limit orders because you are guaranteed to be filled on all your losing trades, but many of your 'potential' winning trades will never be filled.

 

Back to steve....By the time his limit order reached the exchange, price had moved down giving him a better fill. Couldn't price have, just as easily, moved up leaving him behind without a fill?

 

Maybe steve could chime in.

 

~Mike

[Kiwi 16]

  mike21 said:

<<<< Personally I don't like limit orders because you are guaranteed to be filled on all your losing trades, but many of your 'potential' winning trades will never be filled. >>>>

 

I've seen variations of this, including "I don't buy pullbacks, I buy breakouts" based on the same logic.

 

Someone said "Simplicity is the peak of civilization" and such ideas fall into that camp. But someone exceptionally wise said "Everything should be made as simple as possible, but not one bit simpler." And overly simplifying heuristics such as the one above fall into the "one bit too simple" category.

 

As a trader you need to evaluate each possible trade for win rate, win/loss, variability in those, and thus expectancy and profit factor and reliability. You might find that the cost of paying the sometimes very large spread outweighed the number of trades you missed. Or you might not. But unless one does the numbers one doesn't ever know.

 

 

 

 

ps. My favourite quote on simplicity is Oscar Wilde's "I adore simple pleasures. They are the last refuge of the complex."

[mike21 10]

Kiwi,

 

Another saying... 'There's more than one way to skin a cat.'

 

If you prefer to enter trades with limit orders... great. I don't.

 

The majority of my trades are buying/selling pullbacks. You don't have use limits to trade pullbacks. And guess what... I never miss a winning trade, but by using a buy stop I avoid some losers.

 

Yes I try to keep things simple. Warren Buffett once said "trading is simple but not easy"

I completely agree.

 

~Mike

[joshdance 10]

  mike21 said:

Back to steve....By the time his limit order reached the exchange, price had moved down giving him a better fill. Couldn't price have, just as easily, moved up leaving him behind without a fill?

 

Correct. That is the tradeoff. Apparently steve has found through his experience that he will typically not pick a price such that he will be one of the last ones filled (in other words, price will typically overshoot a bit), so he can afford the chance of doing a software-triggered order this way for the chance at a better price. Depends on the market; steve said, I think, that he trades CL. I used to, and can easily see this being useful in that market. In the market I trade, it is highly liquid and virtually no reason to do something like that.

[joshdance 10]

carltonp or anyone else using IB: have you considered using a different data feed for data only, while still using IB for your order routing (I am guessing you can do this in IB?)?

 

I ask because I can't imagine doing business with a data feed that does not accurately show transactions. I personally couldn't live even with the snapshot way they show transactions; I also knew another trader (a good one) who would say that the high and low prints would sometimes not be correct due to the snapshot being off. Why would you trade with this garbage of a data feed? It's almost as bad as spot forex; no accuracy not only in the timing of the transactions shown, but many transactions not even shown period!?

 

So you can get filled at a price that does not show on your chart as having printed... I know a trader who uses IB who this has happened to quite often... so what about IB makes you stay with them?

[gosu 10]

  joshdance said:

carltonp or anyone else using IB: have you considered using a different data feed for data only, while still using IB for your order routing (I am guessing you can do this in IB?)?

 

I ask because I can't imagine doing business with a data feed that does not accurately show transactions. I personally couldn't live even with the snapshot way they show transactions; I also knew another trader (a good one) who would say that the high and low prints would sometimes not be correct due to the snapshot being off. Why would you trade with this garbage of a data feed? It's almost as bad as spot forex; no accuracy not only in the timing of the transactions shown, but many transactions not even shown period!?

 

So you can get filled at a price that does not show on your chart as having printed... I know a trader who uses IB who this has happened to quite often... so what about IB makes you stay with them?

 

IB is my primary broker and I use TWS to execute trades and have charts running IB data alongside another data feed. For me the test of reliability on executions is, am I getting the price I see on the screen? IOW, if I hit the bid or ask, do I get filled? Invariably the answer is yes. I'm on the ES most of the time, however, and the levels tend to be thick.

[joshdance 10]

  gosu said:

IB is my primary broker and I use TWS to execute trades and have charts running IB data alongside another data feed. For me the test of reliability on executions is, am I getting the price I see on the screen? IOW, if I hit the bid or ask, do I get filled? Invariably the answer is yes. I'm on the ES most of the time, however, and the levels tend to be thick.

 

Maybe I'm spoiled because I have a good broker, but seems to me that this should be pretty much the minimum requirement and only major requirement for any broker, namely, you should get filled at the proper price on a market order. Since you're running another data feed which is reliable, I get this. But some people use IB for data and execution, and given the truly terrible quality of their data, I can't imagine trying to trade with it alone, unless it's a long term swing trader who doesn't care about a few ticks missing from the chart here and there, and who doesn't look at the immediate order flow.

[gosu 10]

  joshdance said:

Maybe I'm spoiled because I have a good broker, but seems to me that this should be pretty much the minimum requirement and only major requirement for any broker, namely, you should get filled at the proper price on a market order. Since you're running another data feed which is reliable, I get this. But some people use IB for data and execution, and given the truly terrible quality of their data, I can't imagine trying to trade with it alone, unless it's a long term swing trader who doesn't care about a few ticks missing from the chart here and there, and who doesn't look at the immediate order flow.

 

I find IB data to be very reliable and have no problems monitoring on charts fed by IB data. I trade stock indexes intraday and the fastest chart I monitor is YM 2m. I have used many data feeds over the years and cost is not an issue for me when it comes to getting quality data. My current monitoring setup with 2 data feeds is a result of years of simplifying what I do. At one time I was subscribed to 4 data providers and running 3 different charting packages concurrently on 3 computers and another charting package for EOD that I had to use with Qcollector. None of the data feeds was IB.

 

For how I trade, only a minimum requirement is necessary and getting a "better" data feed doesn't add to my bottom line. Reliability is the key for me and IB has proved to be very reliable over the years, so much so that I found the costly feeds redundant. I've also simplified my charting down to just one provider running on one computer with APC backup. My current cost for data and charts is under $100 per month.

[mike21 10]

  joshdance said:

carltonp or anyone else using IB: have you considered using a different data feed for data only, while still using IB for your order routing (I am guessing you can do this in IB?)?

 

I ask because I can't imagine doing business with a data feed that does not accurately show transactions. I personally couldn't live even with the snapshot way they show transactions; I also knew another trader (a good one) who would say that the high and low prints would sometimes not be correct due to the snapshot being off. Why would you trade with this garbage of a data feed? It's almost as bad as spot forex; no accuracy not only in the timing of the transactions shown, but many transactions not even shown period!?

 

So you can get filled at a price that does not show on your chart as having printed... I know a trader who uses IB who this has happened to quite often... so what about IB makes you stay with them?

 

joshdance,

 

I think the trader you know may be exaggerating a bit.

 

IB does bundle their data. (I think 20 times a second). But all that means is, if there are transactions with no price movement, they may not show up as tick data. So, for example, If you compare a 100 tick chart with unfiltered data next to a 100 tick chart with IB data, the unfiltered chart will have more candles, BUT time and price will be the same. If you compare 5 minute charts they will be exactly the same.

 

I use ensign for my charting with a TraderBytes data feed, but I also have an IB chart up when trading. (I like the 'drag and drop' feature on IB charts for adjusting stops and limits). I have never seen a price print that didn't show up on the IB chart. I'm not saying it can't happen but I haven't seen it.

 

~Mike

[Kiwi 16]

Actually, Josh's trader was 100% correct.

 

With IB's live feed as you use it with ensign the highs and lows of bars and swings can and surprisingly often do miss the extreme ticks.

 

I tried to get IB to fix this a few years back and they came halfway to the party with a 5 second feed (OHLCVT bars at the end of each 5 second period) that is correct but is also 0 to 5 seconds delayed.

 

With Sierra Chart the two feeds can be combined so that the live feed controls the current price, bid and ask and that the 5 second feed is used to fill in any lost highs and lows. So you get a feed that never lags and is correct. But the software needs to take both feeds and combine them or this won't happen.

 

 

Note: if you are using an ib tws chart to show prices then they use the 5 second feed which is why the data is correct. Feed it into ensign and you'll see a different story.

 

Note 2: on ib charts, drop down to 1minute on something that moves reasonably fast like hsi and watch it. They combine the true and live feeds but their algorithm is slightly screwed. Whereas Sierra Chart correctly combines them and if you get a high or low on a bar it's right IB sometimes shortens a bar soon after the new bar starts (realizes the combination is wrong when new true data comes in) - which should never happen.

Edited December 20, 2011 by Kiwi

[mike21 10]

I don't use IB's data feed for my charts, but since IB's data appears to be correct on IB charts I assumed their data feed was not that bad. Apparently I'm wrong.

 

I'll take Kiwi's word that there are still major problems with IB's data feed.

 

Kiwi, who's data feed do you use?

[joshdance 10]

  mike21 said:

joshdance,

 

I think the trader you know may be exaggerating a bit.

 

Apology accepted ....

[gosu 10]

  Kiwi said:

Actually, Josh's trader was 100% correct.

 

With IB's live feed as you use it with ensign the highs and lows of bars and swings can and surprisingly often do miss the extreme ticks.

 

I tried to get IB to fix this a few years back and they came halfway to the party with a 5 second feed (OHLCVT bars at the end of each 5 second period) that is correct but is also 0 to 5 seconds delayed.

 

With Sierra Chart the two feeds can be combined so that the live feed controls the current price, bid and ask and that the 5 second feed is used to fill in any lost highs and lows. So you get a feed that never lags and is correct. But the software needs to take both feeds and combine them or this won't happen.

 

 

Note: if you are using an ib tws chart to show prices then they use the 5 second feed which is why the data is correct. Feed it into ensign and you'll see a different story.

 

Note 2: on ib charts, drop down to 1minute on something that moves reasonably fast like hsi and watch it. They combine the true and live feeds but their algorithm is slightly screwed. Whereas Sierra Chart correctly combines them and if you get a high or low on a bar it's right IB sometimes shortens a bar soon after the new bar starts (realizes the combination is wrong when new true data comes in) - which should never happen.

 

I am on Ensign and it is true that IB's feed will produce slightly different charts especially in fast markets. I always have another feed side by side and have been comparing IB's feed for years.

 

An example was this morning's spike on 6:56 YM 2m, 7:00 YM 5m bar (bar end, PST). The initial IB chart did not show as deep a spike. A manual refresh updated it.

 

Whether it makes a difference in a person's trading, that's for each person to judge for himself. For me it doesn't make a difference, but then I'm not watching every tick either.