Wall Street Erases the Line Between Its Jocks and Nerds
— Read on www.wsj.com/articles/wall-street-erases-the-line-between-its-jocks-and-nerds-1534564810
Data Science and Machine Learning is somethings we all should be aware of. Attached is a podcast with a course overview I am taking to get started.
Instrutor Bios are below
Data Scientist & Forex Systems Expert
My name is Kirill Eremenko and I am super-psyched that you are reading this!
I teach courses in two distinct Business areas on Udemy: Data Science and Forex Trading. I want you to be confident that I can deliver the best training there is, so below is some of my background in both these fields.
Professionally, I am a Data Science management consultant with over five years of experience in finance, retail, transport and other industries. I was trained by the best analytics mentors at Deloitte Australia and today I leverage Big Data to drive business strategy, revamp customer experience and revolutionize existing operational processes.
From my courses you will straight away notice how I combine my real-life experience and academic background in Physics and Mathematics to deliver professional step-by-step coaching in the space of Data Science. I am also passionate about public speaking, and regularly present on Big Data at leading Australian universities and industry events.
Since 2007 I have been actively involved in the Forex market as a trader as well as running programming courses in MQL4. Forex trading is something I really enjoy, because the Forex market can give you financial, and more importantly – personal freedom.
In my other life I am a Data Scientist – I study numbers to analyze patterns in business processes and human behaviour… Sound familiar? Yep! Coincidentally, I am a big fan of Algorithmic Trading 🙂 EAs, Forex Robots, Indicators, Scripts, MQL4, even java programming for Forex – Love It All!
To sum up, I am absolutely and utterly passionate about both Data Science and Forex Trading and I am looking forward to sharing my passion and knowledge with you!
Hadelin de Ponteves
Hello ! Je m’appelle Hadelin de Ponteves et je suis un data scientist passionné !
Etant particulièrement sensible au domaine de l’éducation, je suis déterminé à y apporter de grandes contributions. J’ai déjà investi beaucoup de mon temps dans la sphère de l’éducation, à étudier et enseigner divers sujets scientifiques.
Aujourd’hui, je suis passionné de data sciences, d’intelligence artificielle et de deep learning. Et je ferai de mon mieux pour vous transmettre mes passions. Car c’est en étant passionné que l’on réussit le mieux dans un domaine, et que l’on est le plus heureux dans notre travail au quotidien.
J’ai acquis beaucoup d’expérience en data sciences. J’ai effectué mes études à l’école Centrale Paris, où j’ai suivi le parcours Data Sciences, en parallèle d’un master de recherche en machine learning à l’Ecole Normale Supérieure. Ma page étudiante s’est enchaînée avec une expérience chez Google où j’ai fait des data sciences pour résoudre des problèmes business. Puis j’ai réalisé que je passais la plupart de mon temps à analyser et je développais petit à petit un besoin de créer. Donc pour nourrir ma créativité, je suis devenu un entrepreneur.
Et justement, mes cours vont tous combiner ces deux dimensions d’analyse et de créativité, grâce auxquelles vous intégrerez toutes les compétences à avoir en data sciences, en les appliquant à des idées créatives.
J’ai hâte de vous retrouver dans mes cours et de partager mes passions avec vous !
Hi ! My name is Hadelin de Ponteves. Always eager to learn, I invested a lot of my time to learning and teaching, covering a wide range of different scientific topics. Today I am passionate about data science, artificial intelligence and deep learning. I will do my very best to convey my passion for data science to you. I have gained diverse experience in this field. I have an engineering master’s degree with a specialisation in data science. I spent one year doing research in machine learning, working on innovative and sexy projects. Then a work experience at Google where I implemented some machine learning models for business analytics. Eventually, I realised I spent most of my time doing analysis and I gradually needed to feed my creativity. So I became an entrepreneur. My courses will combine the two dimensions of analysis and creativity, allowing you to learn all the analytic skills required in data science, by applying it on creative ideas. Looking forward to working together !
Hadelin de Ponteves
Fourteen Financial Firms Invest $66 Million
By JUSTIN BAER WSJ
Fourteen of the world’s biggest financial-services firms bought Perzo Inc., an instant-messaging software company, and formed a new technology company that aims to change the way traders communicate.
Led by Goldman Sachs Group Inc., the consortium invested $66 million in the venture, called Symphony Communication Services Holdings LLC, according to a statement from Symphony.
Symphony in turn acquired Perzo, a two-year-old startup founded by veteran communications-software executive David Gurle. Goldman contributed its in-house messaging developments to the new company, which Mr. Gurle will lead as chief executive.
The deal, announced Wednesday, capped months of negotiations that had widened recently to include additional banks. In addition to Goldman, Bank of America Corp., Bank of New York Mellon Corp., BlackRock Inc., Citadel LLC, Citigroup Inc., Credit Suisse Group AG, Deutsche Bank AG, J.P. Morgan Chase & Co., Jefferies LLC, Maverick Capital Ltd., Morgan Stanley, Nomura Holdings Inc. and Wells Fargo & Co. invested in Symphony.
In the statement, Symphony said it expected that many of the financial firms would be early users of the company’s messaging platform.
“Symphony responds to a pressing need across the industry for better methods of communication and collaboration,” Darren Cohen, global co-head of the Goldman principal-investing arm that spearheaded the talks, said in the statement.
The group’s breadth underlined an industrywide push for software that lets employees trade messages instantly and securely. It also highlights Wall Street’s desire to put pressure on one of its biggest vendors, Bloomberg LP. Bloomberg’s communications services remain a ubiquitous presence on trading floors, and the price the data company charges for its terminal—some $20,000 a year—continues to vex bank executives charged with wringing costs out of their trading businesses.
A Bloomberg spokesman declined to comment.
The deal is also a reunion of sorts for Mr. Gurle, who worked with Goldman and other banks during previous career stops at Microsoft Corp., Thomson Reuters Inc. and Skype. He founded Palo Alto, Calif.-based Perzo in late 2012.
In a blog post on Symphony’s website, Mr. Gurle wrote that the new company’s messaging platform “is intended to be used by some of the most time-conscious firms on the planet who are regularly corresponding high-value information—where a delay of a few seconds can have significant cost implications.”
He wrote that Symphony would be available to all financial firms by mid-2015.
The Wall Street Journal reported last week that the bank group was also in talks with one of Mr. Gurle’s former employers, Thomson Reuters, over ways to integrate their messaging platforms.
On Wednesday, a Thomson Reuters spokesman confirmed the data company had held discussions with Symphony.
The news services of Bloomberg and Thomson Reuters compete with Dow Jones & Co., publisher of The Journal.
Write to Justin Baer at email@example.com
The privacy scandal that shook Bloomberg in May could be coming back to bite it. Today, Markit and Thomson Reuters formally announced their new messaging system for finance professionals, Market Collaboration Services. It seems designed to compete with the chat function on Bloomberg terminals, to which Bloomberg owes part of its dominance as a data provider. The two companies said today that Bank of America Merrill Lynch, Barclays, Citi, Credit Suisse, Deutsche Bank, Goldman Sachs, JP Morgan Chase and Morgan Stanley were all on board.
We’ve written before that Thomson Reuters will have a hard time unseating Bloomberg. But traders, bankers, and other financial services professionals we’ve spoken to over the last few months have raised a number of points that lead us to believe that Thomson Reuters and Markit could be more successful than we thought:
Concerns about snooping and data privacy really shook bankers up. In May, Bloomberg admitted that its reporters had access to information about its customers’ usage of their Bloomberg terminals, and there were complaints that they were using it to write stories. Though the fury may have faded, the message has not; third-party technology can pose a threat to the secrecy of the firm. The Markit/Thomson Reuters offering was created in close collaboration with banks and is more customizable, so it may enjoy a certain level of trust.
Not everyone needs a Bloomberg. Bloomberg terminals cost around $20,000 per year, something Wall Street has long seen as a necessary evil. But maybe no longer. “For some big banks, it’s an incredibly expensive instant messaging device,” an executive at one market infrastructure company told the Financial Times (paywall). “They’re saying, ‘we’re spending $120m a year on Bloomberg. That needs to come down’.”
Sharing is caring… about costs. Major banks have already made the decision that employees can share a terminal in some cases, and used the savings to buy cheaper plans from Thomson Reuters that can be customized to fit an employee’s role. A commodities trader, the thinking goes, doesn’t need all the same tools a banker advising on tech mergers does. By contrast, Bloomberg terminals are one-size-fits-all; if you buy a terminal, you have to take all the features it offers even if you don’t need them.
This is already happening; one banker who was not authorized to speak on his bank’s behalf said his team had seen its number of Bloomberg terminals cut down to one, replaced by Thomson Reuters Eikon terminals. The team shares the remaining Bloomberg terminal.
A stand-alone chat function makes a lot of sense. In an email, Thomson Reuters said it “aims to create the largest financial markets messaging community and remove barriers to cross-market communication.” This means installing the messaging service on as many machines as possible, even ones that don’t even receive data feeds. Therefore, employees across the business could have access to the secure chat feature. If fewer bankers have their own Bloomberg terminals, they will need an alternative chat service to communicate with those colleagues that don’t have them.
Clearly, this isn’t a transition that will happen overnight. But with cost pressures mounting and reception already warm, Markit and Thomson Reuters seem to have a better shot at taking on Bloomberg than you might think.
Bloomberg declined to comment.
The 3 Ms of Risk Management
Recent market event have pointed to increasing volatility. As has happened all too many times in the past, risk management disasters continue to plague the industry and show up on the front page of the newspapers. Given the potential for these disasters to occur, lets discuss some required risk management capabilities.
Consider the following scenarios:
A major market move occurs. The Chief Risk Officer (“CRO”) of a broker/dealer wants to know right now what effect this has on the firm. Are they better or worse off? What actions should be taken? To determine the best course of action the CRO needs to know real-time what the positions are and the potential P&L effect. The CRO must also be able to perform a risk analysis immediately. As we have seen, inability to do this will consume resources, raise the firm’s risk profile and possibly cause losses as a result of the market move.
A major firm announces a significant profit restatement. The CRO of a major retail brokerage wants to know which accounts will be affected the most. The CRO needs to know real-time which accounts have concentrations in that industry and SIC code. Since it isn’t clear yet what the full effect of the restatement will be on security prices, can the CRO perform a what-if analysis on specific accounts to determine the potential effects on the firm and the accounts so that proper actions can be taken? Inability to do this real-time will consume resources and increase the risk profile of the firm and the accounts.
In both cases, could the CRO have set up early warnings so that the risk systems would have generated alerts as to problem positions or accounts when specific actions occur so that the CRO can spend less time finding risks and more time managing risks?
The industry has spent many dollars effecting comprehensive risk management capabilities. Ultimately risk management is, however, a process that requires tools and the right mindset, not just a system that measures risk. The purpose of risk management is the following: minimize the probability that an error occurs AND that it goes unnoticed. To do that, a firm must have all the components of effective risk management. The firm must have the ability to perform the 3 Ms of Risk Management: Measurement, Monitoring and Management of risk. In this paper, we will outline the basic capabilities of each of the three areas.
All firms must have the capability to measure their risks. Most firms have risk measurement systems. However, there is a lot more to it than that. Risk measurement involves ALL aspects of the capability to measure risk, not just having systems. To measure risk effectively and accurately, the firm must have accurate and timely information as to its positions, its counterparties and all relevant information regarding its positions and counterparties. This information should ideally be available on a real-time basis as markets move very rapidly and soon the risk analysis may no longer be valid. Measuring this information solely on an overnight (or end of day) basis will not be sufficient as market conditions change during the day, and customer and counterparty activity changes the firm’s risk profile constantly. It is also not sufficient to simply do this several times a day. As market conditions change, the value of the firm’s and its customers’ positions changes accordingly, either favorably or unfavorably. In addition, as customers do trades during the day the firm must be able to track its customers’ accounts as they transact business. This information must be accurate, accessible in a timely manner and able to be retrieved from the firm’s computers and sent to the relevant analytical models for risk evaluation. During some recent risk events, many firms learned that they could not do this effectively, much to their dismay.
So what does effective risk measurement entail? Several key components are required:
The risk measurement methodologies used must accurately measure the risks. There are many different ways to measure risks and firms use most of them. The two basic necessities for a risk measurement methodology to be effective are that they must reflect the risks they measure and all relevant parties must understand them.
Different kinds of firms will require different kinds of risk measures. The measures needed for a portfolio-based approach to risk measurement are not exactly the ones needed for a retail operation. The portfolio approach requires position, position attribute, counterparty and counterparty attribute data. A retail operation will require all that and extensive information at the account level so it can see what individual accounts are doing real-time.
The firm must be able to examine its risks at any level and aggregate up or drill down to any desired degree. For example, a broker dealer must be able to measure risk by security, security type, counterparty and type, industry or SIC classification, currency, geographical location, etc. The B/D should then be able to aggregate up or drill down in any direction (for example by country by currency or vice versa). A retail brokerage should be able to measure risk by account, by account type, by industry, SIC code, etc, and aggregate up. They should also be able to drill down to the account level after starting with a portfolio approach. In addition, a retail brokerage needs to perform sophisticated margin calculations on a wide variety of products. Also, they would need to be alerted when specific activities occur in selected accounts, e.g., large trades or prohibited activities.
The firm must be able to perform scenario and what if analysis on a real time basis for any of its risk measurement categories. For a retail operation, this means even at the account level.
The analytical models used to measure risk must be accurate and measure the right risks. The models must be appropriate to the business and the products. Different products may require different kinds of models and there is nothing wrong with that. Use as many models as is necessary and no more.
The inputs to the models must be accurate. Many firms have a problem with their data and getting it to the right system at the right time. The data must be accurate, clean, and timely. This applies to model-generated data (including the results of risk analysis) as well as historical market data. Without accurate inputs, the model will give misleading results, leading to inaccurate decision-making.
The connections between the systems must be accurate. Feeder systems must feed the inputs to the risk model on a timely and accurate basis, just as the risk system must feed other systems in the same manner.
The systems must work automatically. You should not have to do anything extra for the system to be measuring risk accurately and timely.
The firm should periodically assess its systems and their ability to perform, effecting updated capabilities when necessary.
For effective risk management to take place, risks must be monitored. A firm that simply measures risk three thousand ways but does not monitor it on a timely basis will likely suffer at some point. Risk monitoring includes all aspects of ensuring that accurate risk measurement information is available to the right people on a timely and accurate basis. What does effective risk monitoring entail? Several key capabilities are required.
The firm must have timely and accurate risk information available to the right people at the right time.
The firm must have a set of comprehensive risk reports generated during the day. The reason that a set of reports is necessary is that different levels of management require different levels of risk information. The key criterion is that the reports reflect the degree of granularity and breadth of information required to optimize the decision-making capabilities of the party that gets the reports.
The firm must also have this information available on a real-time basis. This means that it must be available online for the parties that require it so they can see what is going on at all times. The same issues of granularity and breadth apply here.
The risk systems should have the capability to alert the proper manager when preset conditions occur so that proactive risk management can occur. The firm should set up a variety (as many as needed) of risk conditions that different managers are concerned with. These conditions should also be set in a variety of ways. The parties could set up criteria that will generate alerts. The relevant manager could then drill down into the alert to investigate further. The proper action could be taken.
For example, a B/D could set these alerts to show limit utilization above a certain level (e.g., 75%) and by security, currency, counterparty, trading ledger, industry or geographic location. The system alerts the appropriate level(s) of management when the condition is met. The alerts should also happen as a result of a what-if or scenario analysis, alerting the appropriate party to what could happen under certain conditions. For example, an alert could occur if a major market move would cause an X% loss in a particular security. Management can then examine the alert and take appropriate action, if any. These alerts are set by management and should reflect the conditions with which management is currently concerned.
For a retail operation, this would include all the above. It would also need to include alerts at the account level such as a big trade or an account that is utilizing an increasing portion of its credit and is heading toward a potential margin call. For example, an alert could occur if an X% market move would cause a margin call in a large (or small) account(s). Management can set up appropriate conditions for accounts it wishes to monitor and be alerted when those conditions are met. Management can then examine further and take the appropriate action, if any.
In addition to all the reports and alerts, managers must effectively communicate with each other so that they are aware of current conditions.
The first two steps in the process provide the analytics and the tools that managers at all levels must have in order to make effective decisions regarding risks. The final step in the process may be the simplest to explain. After all the risks that can be measured are monitored (those that can be measured. Not all risks can be measured and you should not try!), and after the correct monitoring systems and procedures are in place, the final step in the process is actually managing the risk. This simply means management decision-making when called for, based on the information that is available. Managers at every level must be ready to take appropriate actions when a condition exists that warrants attention. This means proactive actions. Remember that not every risk condition or situation requires action. It possible that, for example, that a limit is exceeded on a trading floor and management becomes aware of it. After reviewing the excess, determining the cause and discussing the possible harm, the appropriate managers may let it stand and take no action. Or, an alert can be generated on a specific account. After drill down and review, management decides no action is called for.
Some of the critical aspects of managing risk effectively are:
The proper analytical tools must be used so that the information to decide possible courses of action is reliable
The proper risk monitoring capabilities, including alerting capabilities that provide this information on a real time basis, must be in place
A risk-oriented mindset must exist in all employees. Senior management must drive this mindset from the top down. Everyone bears some of the responsibility, not just management and risk managers
A willingness to be proactive regarding risk management, treating risk management as a business partner, not simply part of a compliance function
As we all know, there are many crucial aspects to implementing effective risk management capabilities at a firm. It is critical that each phase be implemented at any firm that wishes to effectively manage its risks. This can be summarized relatively simply. The tools for measuring risk must be accurate as must be the inputs to those tools. This means models must be accurate. Data must be clean. The technology behind the system should help the risk management process by identifying risk so that managers gave increased resources for managing risks. Real-time capability is required; batch processes won’t cut it any more. The outputs of the risk measurement process must be available on a real time basis so that managers understand what is happening as it is happening and can take appropriate action. This means everyone gets the info when they need it. Systems that inform management of current conditions go a long way to helping the process. Finally, everyone should consider risk management as part of his or her job. Effective risk management is possible when these conditions are met.
Some Comments on Recent Exchange Rate Activity
The FX markets are critical to smooth functioning markets. Sooner or later EVERY piece of international trade will involve a foreign exchange transaction. That is one reason the FX market is by far the largest market in the world. Another factor is the amount of currency speculation that occurs. Here we simply have those buying and selling solely to try to benefit from anticipated price movement. Finally we have hedging activities, those taking offsetting positions to reduce the overall volatility in a firm or trader’s P&L. These are just some of the factors that affect the FX market, which includes currencies and derivatives.
Some recent actions point to the possibility of increasing exchange rate volatility in the near future.
- Japan has recently begun a change in their macroeconomic management of the Japanese economy. Their desire to cure the deflation which has been hurting the Japanese economy is likely to cause an increase in the inflation rate. This can cause a weakening of the Yen relative to other currencies
- Thailand is considering capital controls and interest rate actions to cool the rise in the Baht
- The US has had very low interest rates for several years as the Fed has been trying to manage the economy back to health. At some point interest rates in the US are likely to rise from their current levels. This will likely cause some increase in exchange rate volatility.
- Problems with the Euro have been plaguing the world, although based on member country interest rates, it appears that markets have calmed down a great deal
While many factors affect currency markets, here is a quick overview of three key relationships affecting exchange rates. We are describing each factor independently even though they are interactive and NOT the only factors affecting exchange rates.
- Interest rate differentials between two countries can affect exchange rates by making investments in the higher interest rate country relatively attractive. This comes from two potential sources. First, the higher interest rate can potentially offer a more attractive rate of return. Second, if the rate is believed to be ‘high’ and likely to come down, there will be a potential capital gain earned should rates decrease. This can translate into an enhanced rate of return for the investor.
- Inflation rate differentials can affect exchange rate by causing a devaluing of one currency relative to another. Generally speaking, the exchange rate between two currencies will depreciate relative to the difference in the inflation rates between the currencies. Inflation tends to weaken a currency and so we could expect the inflation rate differentials to drive a wedge into the exchange rate
- The Fischer effect (named after Irving Fisher) states the nominal rate of interest is related to the sum of the real rate of interest and the expected inflation rate (while this is not literally the relationship, it is close enough for what we are discussing here). As inflation rates rise nominal interest rates should tend to rise with them, although there is often a time lag. If inflation rates rise (as some think likely) expect nominal rates to rise.
In general, an increase in the level of rates tends to raise the measured volatility of those rates. In other words, volatility tends to be higher as the LEVEL of rates gets higher. So if rates go up in the near future (for any reason – inflation, commodity prices, etc) we can expect a corresponding increase in exchange rate volatility.