I think the times of large players doing unsofisticated executions are long gone. Even the time of unsofisticated algos are gone (see below article). And, as @bone said, there's so much hedging and cross activity that figuring out the intent is impossible. Add to that that whales most of the time don't have the urgency to get in and out because they can use their size if necessary. I don't have access to whales, so just my opinion. However, it doesn't mean that all these things(mentioned in title) are useless, I just think you are not dealing with whales, but with sharks, tunas, and sardines. September 23, 2019 New and Improved Algorithms Empower the Buy Side Larry Tabb TABB Group Follow | Profile | More Share Little more than a decade after early trading algorithms transformed the markets, algos account for nearly half of US institutional equity buy-side trading volume. But technology has turned over again and a new generation of algos is emerging. Brokers are rethinking electronic execution and reinvesting to create models that improve performance in ways that previously were impossible. While computer-based trading technology dates back to the early 1970s, modern algorithmic trading was kick-started by the seminal paper, “Optimal Execution of Portfolio Transactions,” by published in 1999 by mathematicians Robert Almgren and Neil Chriss. Their work considered how to execute portfolio-driven transactions that would reduce volatility and lower costs, leveraging computer algorithms to define efficient trading strategies. By 2007, algorithms known by names such as Sniper and Guerilla had become prominent in securities trading, and algorithmic trading became a major factor in how exchanges and markets operated. Another wave of algos was launched in the early 2010s, including strategies such as Dagger, Sonar and Stealth. These algos evolved past simple time and price averaging to take advantage of the increasing sophistication of dark liquidity and new exchange and ATS order types, as well as the increasing amount of available data and analytics. While these algorithmic tools were revolutionary, technology has turned over again, pushing firms to rethink electronic execution and reinvest to create models that improve performance in ways that previously were impossible. Today, trading algorithms are critical to institutional equity traders. Not only do they account for nearly half of US institutional equity buy-side trading volume, they also are the most important aspect of the brokerage business, surpassing high-touch sales trading services and block liquidity in driving buy-side order flow (see Exhibits 1 and 2, below). Exhibit 1: Buy-Side Order Flow Routing by Channel Source: TABB Group Exhibit 2: Buy-Side Order Flow Priorities Source: TABB Group As a result, brokers are re-investing in their infrastructure and trying to leverage new technologies and data science to make their trading algorithms better, more efficient and more effective. New trading technologies, combined with new machine learning and automated intelligence tools, have enabled brokers and algo providers to recalibrate and retarget their strategies as they begin to launch the next generation of algorithms. These new and upgraded algos are being developed to meet stricter performance metrics and benchmark standards, and algorithmic trading is now poised to make a great leap forward in complexity and sophistication. In addition, trading algorithms are increasingly customizable, which enable these newer algos – in conjunction with high-frequency trading capabilities, advanced communications and more sophisticated analytics – to support almost individually defined execution strategies. Overall, the benefits of next-gen algos include transparency and better relationship management for traders with their clients, as well as a more competitive posture and flexible nature in how they interact with the market. With a transparent next-gen algo, a user can see his performance in real time and adjust his trading protocol for greater efficiency and better performance. That transparency, and being able to show that performance to clients, makes the foundation of a trader’s relationships with his clients much stronger. It also may enable clients to work more collaboratively with their traders or brokers as a result. Today, next-gen algos provide the markets with flexibility, visibility and sophistication that just didn’t exist in the first wave of modern trading algorithms inspired by Almgren and Chriss in 1999. The evolution of algorithms and the newest generation of strategies empower users; algo providers’ clients aren’t likely to settle for models with fixed, inflexible structures, which once were all that was available.
Ok, you're describing just the arbitrage between cash and futures, right? I didn't quite get the bottom bit about 'large specs' - it sounds like you're saying large specs want to prevent an arb but I know that's not what you mean so ... The other issue here I'm not sure of is - what's leading? can't cash lead at one time and futures at another (e.g., over releases or volatile periods)...
I don't think it's a static thing. When things are volatile & there's a sense of urgency a large company will not use the C64 SubDagger Geometric Tuner Algorithm but just try get their shit done so "unsophisticated" execution comes back. If we go through longs periods where everything is calm, prices are stable and in line, relationships are clear, then yes there's no urgency to get anything done and so if you have a large order you'll go ahead and work it through a subAlpha TunaEater version 6.89a algorithm and so, depending on the 'technical landscape' (liquidity, retail volume, how spread the product is), that will not be tradeable. I'm not sure what you mean by "they don't have urgency to get in and out because they can use their size." Their size is a downside, not an upside, in a liquid, efficient market-where are they going to push it to so as to get more liquidity? I don't think they're useless as a blanket statement. I like that a lot of these 'auction theories' attempt to come up for some logical explanations as to why price behaves as it does, rather than start with rationalistic assumptions and deduct from there (although I still find a lot of these order flow theories do, they're better than the kind of rationalistic, deductionist logic of 'world moves in patterns. some patterns are fibs. therefore all markets move around fibs."). I think the auction theory/order flow theory type principle are useless sometimes and other times they're not. It depends on the time, the product, and what's going on.
(1) Arbitrage happens if the ES aggressively trades outside of regular spread relative to SPX. Example: Sellers causing arbitrage to happen >> sellers on globex are causing removal of liquidity in cash market! Pushes SPX down. (2) Large Spec are managing aggressiveness of their orders, so managing ES/SPX spread. (prevents arb). (3) They take turns leading. Basically, aggressive buyers will cause arbitrage if they like the price of ES. They cause arb until they don't like the price anymore! This is the PREM (transformed using relative volatility). You can see it often leads price moves. Generally, it shows large spec buying as price moves low, and shows selling as price moves high. This is November 20, 2019
I understand what you mean, but I don't think in today's markets you can simply send huge market order to sell. Much more effective to use some aggressive algo. Not to mention there is much less liability to use someone's pre-approved algo than let some trader make discretionary decisions. Even with algos, people have fat finger situations where they input the wrong values and cause instability. That's why I said "most of the time." When you have size, you can move markets pretty easily as evident by the whole spoofing saga. But yes, in a fast moving markets it's not as easy. Then again, in fast moving markets, market profile is not going to be very useful either. That is true for everything trading related, isn't it?
Sure, true in theory. It's rare but when shit hits the fan there's definitely more primitive execution. Even if you have fast-paced, aggressive execution algos I think the same principle applies: urgency = easier to pick up on intent of the participants and trade it based off order flow. OK, we were talking about different things then. When dead and illiquid, sure. When super efficient and thick, I just wonder is it worth trying to push anything at all? No, e.g., fibs, astrology, candlestick patterns, or anything deduced from a random idea as opposed to something like order flow where there are specific principles induced from observation and understanding of participants and so you can understand why they may or may not apply in a given situation (e.g., super thick heavily spread market where not much is going on - probably not a place to be trying to trade 'order flow' since it's unreadable for reasons discussed previously).
Alright, I think I get this, you're saying execution algos are taking into account he spread so they'll step up or step down the aggression based on where the futures-cash spread is at... Got it, but then assuming most of this lead/follow happens on such a short-time frame (an arbitrageurs time frame) so that you might as well just focus on the price action in one since it contains that information anyway... It's not like you can watch cash and trade futures based off cash since all that is algo'd up to perfection, right? PREM? Huh, never heard of (transformation using relative volatility) but that's interesting. I was thinking there's probably heaps of spreads I'd like to look at simultaneously but they move differently and so hard to tell if one has moved way more than the other or not - this would fix that... nice! Thanks for sharing all of this!
The PREM for an index is the following: add the futures and equivalent ETF then subtract index. 10*SPY + ES - 2*SPX 40*QQQ + NQ - 2*NDX 100*DIA + YM - 200*DJX 10*IWM + RTY - 2*RUT No, the CBOE cash index will lag the tradable instruments. The reasons are complicated. (HFT allows GLOBEX to reach outside listed markets!) It has to do with HFT using synthetic portfolios, and dark pool liquidity to capture the index arb. You have to actually analyse the PREM itself. It has to be charted and smoothed, spread against indicators of itself and so on. It's quant stuff. Retail is not prepared to do this kind of thing. Arbitrage is happening almost all the time. This is because of the inverse relationship to bond market. The bond market is moving the index. For example, if a few hundred million in bonds start to sell off, then the money will go into index futures (so fast that it causes PREM to move and arbitrage to occur). If bond market starts to rally, then they start pulling money out of cash equities using ES (causes arb). Basically ES is supposed to be the hedge for cash stock owners, but they sell ES and buy bonds at the same time.............. (and in reverse) The bond market is KING........