Slippage Testing

Does anyone have any insight into which algorithm from IB is best? I was looking at TWAP / VWAP from IB, Jefferies and CSFB.

The csfb and jefferies suites of algos aren’t IBs. I think they license them or something, but the result is that fees are way higher using those, and from what I’ve heard the results aren’t different enough from the IB suite to justify using them.

Within the IB algos twap/vwap are basically just variations of each other. vwap spreads the order over the entire allotted time period and twap fills as quickly as possible within the % of volume constraint you set. Hypothetically if the order you try to fill is larger than the % of volume you set for the order period then they will behave identically. So take your pick, twap generally if you’re in a hurry, vwap if you want smoother execution over the course of the order.

I generally recommend vwap set to end about 15 min before market close, “allow trading past end time”, “attempt to never take liquidity”, and whatever “max percentage” of volume that you like within reason. In my experience this is a consistent/predictable approach that significantly reduces trading costs and achieves very low slippage relative to daily vwap.

Andrew,

Thank you for your input on this: and for all of your previous posts regarding IB and trading.

I assume you use tiered commissions? So potential for no commissions (or low) on many of your trades when you add liquidity?

On (roughly) what percentage of your trades are you able to add liquidity and reduce commissions?

I have just switched some money to Fidelity which is probably not in the league what you do, but perhaps, better than Folio Investing. Maybe I need to go right to IB. And IB might be easier as I already split up some of my trades during the day.

Much appreciated!

-Jim

I generally use vwap with settings as described above and trade primarily in micro and small cap stocks. So I’m often trading a substantial percentage of the adv. And even doing that it fills around 90% adding liquidity and 10% taking. I do get lots of partially filled orders though, so I just build that in to my trading approach.

I pay about $0.0014/share when adding and $0.0026/share when taking liquidity. Fill prices vs any benchmark are also consistently a few bp better when adding vs taking, so I think its a win-win.

IB has a learning curve, but for anyone actively trading I think it is both the best AND cheapest platform available. Things like slippage and commissions aren’t too important to more passive investors, but with high turnover strategies and/or low liquidity stocks I think its worthwhile to figure it out.

I tried the Accumulate Distribute Algo today and something very weird happened. I wanted to sell 2 shares every 10 minutes for a total of 58 shares (high share price). It immediately filled the entire order on a dark pool. Any idea why that would be?

Its been a long time since I used accumulate/distribute, so I wont guess. Most order types and algo stuff primarily operates in lots of 100 shares, so trying to trade 2 shares at a time may not be supported. Most of these things are also needless complexity if orders are small enough that there is sufficient liquidity to just place a limit order.

When I switched from fixed to tiered pricing a few months ago, the amount of commission I pay instantly dropped around 50%. I almost exclusively trade using VWAP orders, trying to add liquidity as much as possible. I’m not sure how often I add liquidity, but I’d guess at least 60-70% of the shares traded.