The full title of this book is “The Prop Trader’s Chronicles: Short-Term Proprietary Trading Strategies for Both Bull and Bear Markets” by Francis James Chan.
In this book, Chan narrates his story of how he first got started at Swift Trade Securities at Toronto, learnt the art of high volume scalping of U.S. stocks (with tape reading and destination selection as foundation), and his transition to trading at home with CTNY, a U.S. prop trading firm. The book biography wrote that he currently lives in Hong Kong where he trades U.S. stocks for Delta Capital Group and serves as managing director for Atrium Asset Management.
For readers familiar with SMB Capital, Chan traded in a very similar fashion. The main difference I see is that while Chan scalped for 2-3 pennies in a market maker style, SMB Capital traders go for longer swings until there is a reason to sell. In terms of trading platforms, Chan used an in-house Propser Pro platform at Swift Trade, the Sterling Trader Pro software at CTNY, while SMB Capital uses the Lightspeed platform.
I learnt two things from this book. One, the interesting fact that the bid-ask spread cannot be less than one penny only applies to public venues and not dark pools. That explains why I always see trades occurring in-between the bid-ask spread. I had previously thought that they were simply large block orders and the prices reflected the average price.
The second thing I learnt is to not be too concerned about extensively testing all strategies with large sample sizes. If you can find a transient edge that you can make money off, that’s fine too. Don’t be too concerned with finding an edge that works in perpetuity, as the key to survival is to be able to constantly adapt to the changing market conditions.
Chan’s point on using the S&P 500 futures time & sales as a market sentiment indicator also clicked with me as I have just started to also look at the intraday price action of S&P to better time my entries and exits for stocks.
All in all, this book is a quick read and is a good introduction for those unfamiliar with intraday stock trading strategies used by prop trading firms.
Traditional ECN Pricing Scheme and Broker Commissions
- ECNs charge a fee for removing liquidity, and pay a rebate for adding liquidity.
- For any of the major displayed venues (ARCA, NASDAQ, BATS), the ECN would typically pay you a rebate of something in the neighborhood of $2.00 to $2.50 per thousand shares.
Why An ECN Inverted Pricing Scheme Can Work
- Direct Edge exchange’s EDGA system used to pay a rebate for removing liquidity, and charge a fee for adding liquidity.
- When buyers want to aggressively buy at the market, the first destination they hit will be EDGA because they get paid for removing liquidity, compared to other destinations where they have to pay a fee to remove liquidity.
- Sellers will add liquidity on EDGA because their orders will likely to be filled before everyone else’s at the same price. They might get their sell limit order filled at the national best offer right before the market absorbs the bid level and takes the price down.
Examples of Routing Instructions to ECNs
- Specifies where the order flows to in the event the order isn’t matched.
- Island ECN – DOT I: NASDAQ book >> NYSE floor
- Island ECN – DOT D: NYSE floor
- Direct Edge – ROUC: EDGA / EDGX >> Dark pools >> NASDAQ OMX BX >> NYSE floor >> EDGX
- NASDAQ ECN
- Code: Q
- Island and Instinet merged in 2002 to become INET. Bought over by NASDAQ in 2005.
- NYSE ARCA
- Code: P
- Archipelago merged with Pacific Stock Exchange in 2000. Bought over by NYSE in 2006.
- Direct Edge’s EDGA
- Code: J
- Originally free, then used an inverted pricing model, now a normal deep discount ECN. Owned by Knight Capital Group.
- Direct Edge’s EDGX
- Code: K
- Traditional ECN. Owned by Knight Capital Group.
- BATS BZX
- Code: Z
- BATS BYX
- Code: Y
- Originally the International Stock Exchange (ISE), bought by BATS. Inverted maker-taker model.
- CBOE Stock Exchange
- Code: W
- NASDAQ OMX BX
- Code: B
- Originally the Boston Stock Exchange. Inverted maker-taker pricing model.
- NASDAQ OMX PSX
- Code: X
- Originally the Philadelphia Stock Exchange. Offers price-size priority order matching rather than the traditional price-time priority (i.e. larger size has greater priority at the same price level).
Time & Sales Reports All Trades
- Every quote and trade on the U.S. equity exchanges is consolidated by the Consolidated Tape Association (including trades executed at Dark Pools).
No Order Can Be Filled Outside the NBBO
- The trade-through rule in Regulation NMS (National Market System) requires that no ECN, exchange, or other destination can fill an order outside of the current National Best Bid and Offer (NBBO).
- All the displayed liquidity on the NBBO, if not cancelled, must be matched by removers’ orders before the price level will actually break up or down.
Trades Can Take Place In-Between the NBBO
- Regulations specify that the displayed NBBO on visible venues must not be tighter than one penny for stocks priced above $1. However, this does not apply for hidden non-displayed or dark liquidity.
- Trading the midpoint of the NBBO has become a major asset for those who can maneuver with it and can, by itself, be something of a soft edge in the markets or even contribute as a component of a hard edge.
HIGH VOLUME SCALPING
Use Time & Sales of S&P 500 Futures as Market Sentiment Indicator
- The Time & Sales of the S&P 500 index futures contract was an excellent indicator of market sentiment.
- It can be used as the last-step confirmation signal for trades.
High Volume Scalping Model
- Exit at between breakeven to a maximum loss of two pennies.
- The typical profitable day would look like a series of breakevens (some of which are net positive after costs with ECN rebates), some winning trades of one penny each, an occasional winning trade of three or more, and a few losing trades of one or two.
- Stop loss orders were never used. All loss limiting exits were done manually, especially since there are opportunities to exit using different routing strategies that sometimes cancel out all costs and occasionally come out with a profit on a breakeven trade.
- Enter passively with orders that add liquidity and then manage the resulting positions by exiting aggressively.
- Enter aggressively when the odds look good and then try to either exit passively by adding liquidity or exit aggressively by removing on a well-priced ECN.
- Layer orders around a range of price levels to continually collect rebates for adding liquidity as the net position continually changes when their orders are filled. Depending on the average daily range on a stock, and your current maximum trade size, this strategy and its variations are some of the most time tested methods of scalping with ECN rebates used directly in maintaining a positive edge.
- High-volume low-priced stocks were popular for slow-moving but easy-to-read movements to scalp one penny at a time, such as Citigroup between 2008 and 2011.
- High volume but higher priced stocks tend to be popular for scalpers who aim for more than one penny at a time, such as General Electric.
Reduce Risk by Scaling In and Out
- By trading around a position, scaling in and out of massive positions, one can constantly reduce each trade’s risk to less than the size of his average profit-taking size.
Scalping Risk Model for Trainees
- Trainees start off with a maximum loss of $25 per day (at 100 shares, that amounted to 10 to 25 consecutive losses in the style I traded — which typically meant you either had absolutely no idea how to read the tape, or you were too stubborn to admit you were wrong).
- Swifties would begin with one lot (100 shares, which comes to $1 per penny of price change) per trade. This number, of course, quickly rose for trainees who achieved consistent profitability at each level.
- Psychologically, many trainees face a mental block around 5 to 10 lots (500 to 1,000 shares) as the emotional effects of the monetary impact of $5 to $10 per penny of price change on the stock begin to take effect.
- As the share size increased, the maximum loss increased to allow for any intraday draw-downs.
- An additional rule stipulated that once a trader has locked in profits during the course of the day, the trader would only be allowed to lose half of those profits in the remainder of the day. This is effectively a trailing stop loss on the traders.
- At the end of each trading day, every trader on the floor would print a record of all the trades made that day to review overnight.
Examples of Dark Pools
- NYFix Millennium ATS
- Credit Suisse Cross Finder
- GETCO Execution Services ATS
- Knight Capital Group’s FAN (seeks both public + dark) and COVERT (only dark)
- Level ATS by eBX LLC (JV of Citigroup, Credit Suisse, Fidelity, Lehman, Merrill Lynch)
- Goldman Sachs Sigma X.
Gaming Dark Pools
- Seek out large orders resting in the dark pool systems and play the market against the large institutions that entered those orders.
- Throw out a few small orders of less than a thousand shares each onto a dark pool to locate any possible blocks of larger orders resting on the system that would indicate a large amount of buying or selling interest.
- If probing buy orders were immediately filled, and it looked like a large seller is in the market on the dark pool system, then begin to load up on a short position by swiping up the visible bids posted opposite the big potential seller’s likely dark offer price.
- Effectively this consumed the liquidity that the large seller would optimally hope would not budge (or at least would move minimally) in the process of waiting for their sell orders to be filled. Other market participants would be more likely to post offers at the same or worse prices as the national best bid had decreased.
- If the market is taken down a few levels, new offers would be posted at lower and lower prices, eventually stepping in front of the prices where he had found the big seller’s dark resting sell orders.
- Eventually the large market participant attempting to sell on the dark pool would be forced to lower its hidden sell limit price, which was often pegged to the visible national best bid and offer on many of these systems at the time.
- The short position can then be unloaded using buy orders on the dark pool route, which were immediately filled by the liquidity provided by the large market participant.
There is No Holy Grail
- There’s no single strategy, no single edge that can be exploited forever.
- There are, however, many types of edges that come and go, and that’s more than enough for a full-time trader to make a living.
- The point is that your job as a trader should be to find and develop an edge that works today, in the market that you will be trading for the coming month or year. And when that begins to weaken, you adapt and develop a new one.
Make Money From Your Hunches, Even with a Small Sample Size
- Markets change and evolve. It’s all well and good if your goal is to develop a trading system that had been back-tested over 10 years of data and then tested on out-sample testing with another 10 years of market data — but even if you had done so, there is still a very good chance that the system will outright fail to deliver one penny of profits tomorrow and for the next 10 or 20 years.
- As a trader, your job is to make money today and tomorrow and for the next 10 years.
- Sometimes a relatively simple idea that just seems to work based on a hunch followed by a little market data from recent weeks pasted into an Excel spreadsheet — as long as it’s attached to a decent risk management technique that protects you from excessive losses — is all you really need to extract more income from the markets.
- Never dismiss these opportunities just because the idea seems too simple, too obvious, or any other criticisms based on fear, uncertainty, and doubt. If you have a hunch that you think might work in the markets, test it on a small sample of data first. If possible, test it on recent data using Excel, which may or may not be feasible depending on the nature of your idea — or in some cases, on a larger sample.
Difference Between Hard Edges and Soft Edges
- A soft edge is one that’s good enough for a trader to make money with enough discipline, consistency, and skill. The real mathematical advantage behind such an edge is rarely more than 1 to 5 percent above a raw 50-50 bet, a very large dose of discipline, consistency, and skill should be developed to properly exploit these soft edges.
- The degree of necessity of hard work and discipline is greater when your strategy’s edge is just a soft edge.
- A hard edge (or transient edge) is an inefficiency discovered in the market, at any given time, that can be exploited by anyone. Hard edges are far rarer in any market — but when they fall into your lap, don’t miss the opportunity to exploit them as much as possible while you still can.
Difference Between Scalping NQ Futures and Stocks
- NQ price levels appeared to move with what could only be described as a choppy hyperactive motion in comparison to the relatively slow and gradual appearance than stocks like GE exhibited at the time when a price level was about to break.
- Watching the Level 2 quotes on the NQ futures contract would instantly inspire a certain level of nervousness in a scalper who was accustomed to the calmer and more gradual pattern of getting smaller, getting smaller, shrinking … and … break on liquid NYSE listed stocks.
Commissions and Profit Split Percentage At Proprietary Trading Firms
- For broker commissions, US$0.001 per share ($1 per 1,000 shares) is considered relatively high for a scalper who trades more than 100 times round trip per day with decent sizes.
- A profitable scalper with a proven track record can typically negotiate a rate of US$0.0002 per share (20 cents per 1,000 shares traded), with any Toronto or Montreal based proprietary trading firm.
- Profit split starts at 35% for beginners (i.e. 35% of trading profits will be paid to the trader), up to 85% for experienced traders with verifiable track records.
Adopt a Trading Style Suited to Your Personality
- If you require absolute confirmation to qualify a concept to be held as a real, hard belief in your mind, then focus more on developing automated trading algorithms rather than manual trading.
- If you tend to accept and believe all concepts that have been suggested to you throughout your entire life, develop manual trading skills.
Keep Adapting and Never Give Up
- And as with relationships of all kinds, no single loss, or even a streak of losses, will mean the end of the world. Sometimes an expected event might naturally weed out the weakest hands in the game, but the survivors are the ones who can adapt and survive to face the challenges of another day.
- Every day, the best we can do is put forward our personal best in our ability to adapt to new challenges and face new forms of adversity. Sometimes, we find the best aspects of the hand we’re dealt simply by being able to turn a disadvantage into a new asset.
Surrogate Market Making
- Similar to high volume scalping, keep a positive edge by earning rebates from ECN systems rather than constantly paying fees.
- To hedge against adverse selection (e.g. only getting orders filled when the market is ready to move against you rather than in your favor), take a position on the opposite side of the market in an instrument (e.g. stock or index ETF) that is partially correlated with the primary instrument. Use historical tick data to find stronger intraday correlations and co-integration figures.
Layered Position Sizing
- Splitting up your entries and exits into smaller orders makes a lot of sense for stocks that move with a very choppy personality on a typical trading day, because it’s unlikely that the stock will stand in place for long and there’s a lot more room to take extra profits by scaling in and out rather than making a flat single-entry single-exit bet.
Opening and Closing Auctions
- Enter an order with OPG time-in-force prior to the cutoff time to participate in the NASDAQ opening auction, and with CLO time-in-force prior to the cutoff time to participate in the closing auction.
Intraday M&A Scalping
- For announced M&A deals, if recent market action showed a strong intraday support at say 12.30 and typically ranges between 12.30 and 12.45 due to time value, the trader can begin scaling into positions around 12.29 to 12.35 and taking profits between 12.42 to 12.44 by adding liquidity on both sides of the trades.
- The primary risk is unexpected announcements that might cause the market to doubt that the deal will hold up.
- Find two correlated and co-integrated stocks and take a long position in one and a short position in the other. You can also go long a underpriced stock and short an overpriced stock based on fundamentals.
- Use tape reading and destination selection to optimize your entries and exits and reduce transaction costs.
- Look for more obscure pairs because the obvious pairs (e.g. Coca-Cola and Pepsi) are already played by a very large pool of algorithms and other pair traders.
Fading Initial Moves in Currency Markets in Summer
- Chan employed this strategy over the summer of 2008. He noticed that major currency pairs involving the U.S. dollar would move within the first half hour from 8am EST when banks in New York start business. There will be another jump in volatility around 9.30am EST when the stock market open.
- In the summer there is a tendency in the currency markets to establish the majority of its range during this period of increased volatility. After the jump, there is often an opportunity to layer a few orders in the opposite direction of the initial moves and catch a relatively predictable 50% retracement of the initial move.
- The reason for this is likely to be a combination of the lack of real volume in the summer months combined with the fact that many of the larger market makers in the interbank foreign exchange market tend to be placed into a counter-trend exposure by those daily moves and must find a way to flatten out their exposure.
- With relatively low volume market environments, their process of flattening out their exposure would push back a fair bit and in some cases even reverse the trend.