IB Adaptive Algo: How Good Is It?

Here is the story:

I have tested side by side for 3 months straight using tdAmeritrade API (market orders) and IB adaptive orders (MKT order with 'Normal' adaptivePriority) to buy a bunch of stocks at the same time and sell them later at the same time.

I have consistently gotten worse pricing (after commission paid) in IB that in TD that's commission free. I also tried IB MKT order straights, doesn’t seem to do better than TD too. I also tried 'Urgent' adaptivePriority, not much difference.

I’m puzzled and surprised as TD is known to get paid for order flow so I would assume IB would get better prices (after commission). But that’s not the case in my experience.

I really want to make IB work but the results are frustrating to me.

I am hoping that I somehow did something obviously wrong and people here can share their experience!

 

That's interesting. At the end of the month what is the ~%yield difference between the two platforms?

Off the top, can you verify that all of the parameters are identical for each platform?
e.g....algo parameters, adaptive order priority, -urgent, -normal, -patient

Are the numbers and types of parameters identical for each platform?

Are the orders sent at the "same time".

 

1). the difference of monthly returns: 0.4% less IB vs. TD. Caveat here: I every now and then used manual intervention on IB side. But I had daily returns for those non-intervention days and my estimate is at least 0.4% or even more for a month.

2). notice that TD doesn't have anything of adaptive algo sort. I used a vanilla TD mkt orders.

3). ib and TD are all fired up using multi-thread Pythons. Supposedly they should be at the 'same time' but I do suspect that IB might have a bit lag as it needs to go to its gateway and then to IB server vs. TD's straight restful API to its server.

 

Wonder about TD and IB mkt orders executed same time.

 

I doubt a rest api is faster than ib connection. Either way, what matters is after the broker gets the order which you don't know

 

Adaptive Algo are not really tailored for stocks. Use midPrice algo instead. There are many several algos available at Ib which could give better fills depending of time of day, duration of your trades etc.

 

Also a big difference, you will notice if You change the route of your trades. (This is more important than algos used for your execution)

 

So what is it tailored for? And does it work there?

 

One reason might be that Ameritrade's affiliated market makers give better fills because they expect Ameritrade's clients' orders to be particularly uninformed, compared to IB's clients?

 

If this is US equities, doesn't reg nms prevent this?