Introducing POV to our Algo Suite

By Suren Markosov, head of Quantitative Research, Anboto Labs

We’re introducing Percentage of Volume (POV) to our robust toolkit of algorithmic strategies. Providing our clients with comprehensive and effective solutions, this addition empowers traders to optimize their cryptocurrency executions.

The POV algo (also known as “Volume Inline ‘’, “Participation Rate”, or “With Volume”) is designed to execute while controlling its share of the real-time market volume at a client-specified rate. The algo aims to control market impact by adjusting to market conditions and allowing the order to follow the natural rhythm of the market. This allows traders to execute more discreetly (with less signaling) by “blending in” with the overall traded volume.

The charts below show how the POV algo’s execution path follows the market volume at different participation rates settings.

The Advantages of POV

As we mentioned earlier, one of the main advantages of the POV algo allows the traders the flexibility to control their orders’ volume participation rate directly. This direct control over the participation rate makes the price impact of the order more predictable. By comparison, schedule-driven algos like VWAP (Volume Weighted Average Price) or TWAP (Time Weighted Average Price) control the participation rate indirectly, usually by adjusting the duration of the order, and rely on historical estimates rather than real-time volumes.

• For example, if the order size is 10% of the typical (e.g. median) historical daily volume in a given coin pair, and the client places a 12-hour TWAP (or VWAP) order, the client can expect their volume participation to be approximately 20%.
• However, if the volume today happens to be significantly less than the typical daily volume, the client’s TWAP order is going to have a higher than 20% participation rate.
• Schedule-driven algos like TWAP or VWAP are able to initiate/create volume in order to complete the order whereas POV algo generally follows the market volume without initiating it (note: volume may still expand in reaction to a POV order if the participation rate is set sufficiently high and the order size is large enough).

One of the disadvantages of the algo is that it does not guarantee complete execution of the order, e.g. if there is no sufficient market volume during the life of the order. For orders where execution completion is more important than control over the participation rate, clients may prefer to use schedule-driven algos like VWAP or TWAP, or slicing algos like Iceberg, which typically guarantee completion (provided the order is within the client’s price limits and other conditions, e.g. being able to cross the bid-ask spread).

What Price to Expect from your POV order

The average execution price of the POV algo is expected to be close to the VWAP over the period of the order’s execution (the so-called “interim VWAP”). Some traders prefer to judge the quality of their POV algo’s execution by another benchmark, called Participation Weighted Price or PWP. The “PWP-X%” (“Participation Weighted Price at X% of the market”) is calculated as VWAP of the period from order start until (Order Size / X%) has traded in the market. The X% in the PWP calculation is typically matched to the participation rate specified by the client in their POV algo.

• So, if the client specifies 10% participation rate on a POV order for 20,000 coins, “PWP-10%” benchmark will be calculated as the VWAP of the first 20,000/10% = 200,000 coins that have traded since the time the client sent the POV order.

If the POV algo stays close to its target participation rate, then the PWP and Interim VWAP benchmarks should be identical. PWP benchmark can be useful in checking for POV algo’s deviations from its client-specified participation rate. PWP is also used by traders for assessing other, non-POV algos, including algos like Implementation Shortfall.

Subtleties of the Participation Rate

The participation rate is the main setting of the POV algo — it sets the aggression/urgency level for this strategy, i.e. the higher the participation rate setting, the more impact the user is willing to have, in order to complete the order quickly. A typical low participation rate is 5%, and the most commonly used highest participation level is around 1/3rd (33.33%).

Note that when using POV at 1/3rd (33.33%) rate, and 10000 coins trade away from our POV algo order, the algo would need to trade approximately 5000 coins (not 3333). This is because the algo needs to be 1/3rd of the total volume of 15000 coins (10000 traded away + 5000 traded by the algo). So, at 33.33% user-specified POV setting, the orders sent by the algo, are actually 50% of the volume that traded away, i.e. “Percent of away volume” = “POV Rate” / ( 1 — “POV Rate”).

The chart below shows the relationship between the client-specified POV Rate and the percentage used by the algo when sending orders in reaction to the market volume that trades away.

The above can cause compounding effects, e.g. if you have 2–3 participants in the market at the same time, trading sizable orders at 33% POV. Thus, when trading a large size order (as a percent of the daily volume), it is advisable to keep the participation rate settings lower (closer to 5–10%) and use price limits, in order to avoid excessive price impact.

One of the mechanisms Anboto’s POV algo uses to mitigate the risk of excessive impact is to allow the POV algo some “breathing room” relative to its participation rate target. In addition to the Participation Rate setting, clients can specify a parameter which determines how far the algo can fall behind its target before it forces itself to catch-up.

• Setting “POV slippage risk” parameter to “high” indicates that the client is willing to allow the algo to fall behind its participation rate target, thus reducing its overall price impact, and allowing for larger percentage of the order to be executed at passive prices (e.g. at best bid for a buy orders, or best ask for a sell).
• Passive execution would have the additional advantages of lower exchange fees (paying maker fee rate instead of taker), allowing for better execution prices.
• This cost advantage needs to be weighed against the higher risk of falling behind if the price starts trending aggressively against the algo.

Alternative Uses of POV algo

Some traders use the POV algo as a substitute for Implementation Shortfall (IS) algo, using the participation rate as a proxy for the IS Algo’s Urgency parameter (e.g. 5% for Low Urgency, 10% for Moderate Urgency, and 20–25% for High Urgency). This can be especially effective if the POV algo is configured (or manually adjusted by the client) to vary its participation rate depending on whether the order is “in-the-money” (lower for buys, higher for sells) relative to the Arrival Price (mid-price at the time of order arrival).

If you have any questions about our POV algo or any other aspects of our product offering, please feel free to reach out to us at info@anboto.xyz