I’ve tested most of the IB algos, including arrival price. You’ll need to ask a more specific question to get a detailed answer, but I had several complaints about arrival price.
The simplest answer is that it is basically a volume participation algo when the price is moving favorably (down if you’re buying or up when selling), but if the price moves against you it slows way down and often doesn’t complete orders. This is especially true for lower volume stocks. I’ve had orders fill <50% when the price was only down .25% from my arrival price even though there was ample and steady volume given my max vol % throttle and “aggressive” setting. It just sucks at filling orders if the price doesn’t move favorably. and the logic it uses isn’t customizable enough to fix it.
It also doesn’t have an option to try and not take liquidity, and in my experience it executed most transactions by submitting marketable orders even when set on “neutral”. Passive did tend to not take liquidity, but it simply did not fill unless the price moved favorably. Maybe not a huge deal for some, but it’s worse that alternatives at keeping commissions low in this regard.
Generally it just makes “odd” choices. I go back and look at the hours in which it was active and it will have elected to not make any trades for a whole 2 hour period when the price was better than arrival (and volume was present), and then suddenly tried to fill 20% of the order in 10 minutes when the price started moving the other way.
I never use the AP algo any more. There are just better alternatives in every situation.
If you’re looking for minimum hassle and a virtually guaranteed reasonable price I recommend VWAP. If you’re not in a hurry and your order is less than 5% of normal volume you can often try to never take liquidity which keeps commissions down around $.001/share, so you’re almost trading for free and still always around or better than the avg price.
If you want to fill as quickly as possible I’d go with Volume participation.
If you’re worried about liquidity I like to use a combination of one of those order types and CSFB Crossfinder, which is often able to fill large chunks of an order very quickly out of dark pools. There is a surprising amount of “large block” liquidity in the dark pools for many small stocks. But do note that the CSFB algos will always charge a >.5 cent per share commission since they don’t use IB smart routing. It sucks they don’t let you do basket trading with multiple algos, but splitting an order over multiple methods doesn’t take too much time if you’re monitoring execution.
If the best possible execution is super important to you I personally still like using the accumulate/distribute interface with a relative limit tied to some short term moving averages and other conditionals - but this method requires a lot of experience, time, and attention to do better than VWAP consistently.