Yeah, that’s it! By screwing IBKR customers with bad fills like piercing the NBBO often when I was making hundreds or a thousand day trades, I took comfort knowing if I was less informed I would get better fills. Makes sense to me now, thank you!
You are using an out of date browser. It may not display this or other websites correctly.
You should upgrade or use an alternative browser.
You should upgrade or use an alternative browser.
IB Adaptive Algo: How Good Is It?
- Thread starter largib
- Start date
So what algo would you suggest for a simple strategy that just wants to get the orders done in, say, less than one minute between at least 3 minutes after market open and 5-10 minutes before market close (thus some sort normal market hours)?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, I had fully wanted to move to IB and expected IB to do better than TD and the results started to puzzle me and thus I delayed the move to further test.
Another thing: I also experimented with IB's vanilla MKT orders and they seemed to perform worse than TD too.
I am willing to further tweak to make IB work (at least no worse than TD) as IB seems to be more for traders than TD (or Etrade or any others). So please help to suggest something (will now start to switch to midprice to see)!
Another thing: I also experimented with IB's vanilla MKT orders and they seemed to perform worse than TD too.
I am willing to further tweak to make IB work (at least no worse than TD) as IB seems to be more for traders than TD (or Etrade or any others). So please help to suggest something (will now start to switch to midprice to see)!
Last edited:
So what is it tailored for? And does it work there?
Adaptive algorithms are mainly used for futures trading. There is a good discussion on the forum by db01 (who is very knowledgeable on the subject) that explains in detail his findings about midprice/adaptive algorithms. In short, adaptive algorithms use time in their calculations of entry points, so they are delayed (as you can see with the different types, such as patient, normal, etc.). My findings show that they are not suitable for use during high volatility as the slippage taken can be significant (for example, in MNQ ES futures). Therefore, my conclusion is that it is best to avoid using them. However, they may be useful for trading illiquid futures, such as rough rice or milk. If You trade stocks try to use Midprices algo. (not supported by futures though)
These were large cap US stocks. Nothing fancyIf this is US equities, doesn't reg nms prevent this?
My understanding: For example, if IB only sends your orders directly to the market without selling the order flow, your order is 100 shares and the order book has 50 shares on it, then you get a fill for 50 and the other 50 at a worst price. Market makers might want to take your entire 100 shares at the best price (if they believe the average Ameritrader is a typical retiree investor instead of an informed laser focussed day trader), even if the order book has only 50 shares on it.If this is US equities, doesn't reg nms prevent this?
Adaptive algorithms are mainly used for futures trading. There is a good discussion on the forum by db01 (who is very knowledgeable on the subject) that explains in detail his findings about midprice/adaptive algorithms. In short, adaptive algorithms use time in their calculations of entry points, so they are delayed (as you can see with the different types, such as patient, normal, etc.). My findings show that they are not suitable for use during high volatility as the slippage taken can be significant (for example, in MNQ ES futures). Therefore, my conclusion is that it is best to avoid using them. However, they may be useful for trading illiquid futures, such as rough rice or milk. If You trade stocks try to use Midprices algo. (not supported by futures though)
Do you have a link for these discussion please?
cannot find it.