![\"-](\"https:\/\/i0.wp.com\/ucanalytics.com\/blogs\/wp-content\/uploads\/2013\/09\/Music.jpg\")
<\/a>– by Roopam<\/p><\/div>\n
Lehman Brothers, a financial services company, filed for bankruptcy on 15th September 2008 – exactly five years from today. A company founded in 1850 disappeared overnight. The chain of events to follow brought the overall financial system across the world to its knees \u2013 hence started the biggest economic crisis since the great depression. A fair blame for the crisis was attributed to the financial models used by the rating agencies, banks and financial services. Now exactly half a decade since the meltdown, it seems a good time to ponder about wise analytical and financial models.<\/p>\n
Being Wise<\/span><\/h2>\nEarly to bed and early to rise, makes a man healthy, wealthy and wise. I must have been four or five when I heard this phrase. Possibly, at that early age I got a sense of what human beings chase as adults – being healthy, being wealthy and being wise. Later when I grew up a bit, I could make out what being healthy and wealthy meant. However, for the longest time I used to think being wise meant someone who could do a math problem really fast. Today, some 30 years since, I am still pondering about the term \u2013 wise. When I see people around me they are all actively chasing health and wealth \u2013 but have they given up on being wise? Before this we must ask, what does being wise mean? I may have an answer and yes, there are some concepts of analytics in here too.<\/p>\n
![\"Laughing](\"https:\/\/i0.wp.com\/ucanalytics.com\/blogs\/wp-content\/uploads\/2013\/09\/laughing-Budha.jpg?resize=225%2C300\")
<\/a>Laughing Buddha – by Roopam<\/p><\/div>\n
Traffic can make the polite use cuss words. What is it about traffic that can perturb us so much? Is it that we want to be somewhere and being stuck is just not acceptable? Awards can make the humble proud. Accolades and appreciation have that kind of influence on all of us. These are just a few examples, in daily life; there are tons of things that ruffle us.<\/p>\n
The Buddha<\/span><\/h2>\nWhen I think of the wise, say Buddha, they say he was unperturbed in most situations. You praised him and he stayed calm. You offended him and he was unflustered \u2013 at least that is what most Buddhist stories say. So was he detached and did not care. Is that the secret of being Buddha? He did leave his family to live the life of ascetic. However, then he did realize that it was useless, stopped running and came back to the world. He was absorbing information, as we do, without getting affected. Rather, he provided solace to the perturbed. Possibly, that is wisdom or being wise.<\/p>\n
I think the secret to wisdom or being wise is self-knowledge. Once you know who you are, the external factors would not perturb your inner core. \u00a0It is certainly not about knowing it all but knowing what you do not know. Knowing one\u2019s own limitations and accepting them makes a human wise.<\/p>\n
The State of Equilibrium<\/span><\/h2>\n![\"Equilibrium\"](\"https:\/\/i0.wp.com\/ucanalytics.com\/blogs\/wp-content\/uploads\/2013\/09\/Equilibrium1.jpg?resize=300%2C135\")
<\/a>Poise, I have always found this word funny. The sounds of some words tickle your eardrums \u2013 poise is one such word for me \u2013 not the meaning of the word just its sound. Poise actually means equilibrium, state of balance or composure. Let us consider the adjacent graph. The green zone in the middle is the state of poise. The state you find yourself while reading a good book on Sunday afternoon, or playing with the kids, or staring at the night sky. A manic-depressive often lives in the red zones. Most of us traverse through all the zones in a day\u2019s time \u2013 the highs and the lows. Traffic throws us down and awards up.<\/p>\nWise models<\/span><\/h2>\n![\"Means\"](\"https:\/\/i0.wp.com\/ucanalytics.com\/blogs\/wp-content\/uploads\/2013\/09\/Means.jpg?resize=362%2C110\" "\\"The")
<\/a>The law of Average<\/p><\/div>\n
Keeping the above in mind, let us come back to analytics. Recall the most fundamental concept of average. Now, consider two different kinds of averages (a.k.a mean) one with million data points and other with just five. Both have the same value of say 100. Now, you throw in some new information and add 1000 to both these values. The first one will stay mostly unaffected but the second one will jump up 150%. The down side or adding -1000 will have the same impact on the values.<\/p>\n
Regression towards the Mean<\/span><\/h2>\nThe following is an example of regression model. The relationship between age of a teenager (in years) and his height (in feet) is displayed in the model. This is most certainly not the best of models but is used for the purpose of illustration.<\/p>\n
![\"Height=0.1\times](\"https:\/\/s0.wp.com\/latex.php?latex=Height%3D0.1%5Ctimes+Age%2B4%2B%5Cepsilon++&bg=ffffff&fg=000&s=1&c=20201002\")
<\/p>\n
Now the beta values or 0.1 and 4 in the above case are the soul of regression models. The beta values are actually average or mean values of the parameters. The ideas we discussed above about averages applies to the beta values as well. Notice, this is not a universal model. It will only hold true for teenagers. For a 100-year-old male, the model will not estimate him to be 14 feet tall.\u00a0The same applies to logistic regression discussed in a previous article<\/a>. Genuine models have well defined boundaries for which they will hold true. They don’t pretend to be true universally.<\/p>\nWe have already created the foremost requirement for wise models. They do not fluctuate much when some unusual information comes their way.\u00a0On the contrary, wise models understand the limitations of their knowledge and define the boundary for which they will hold true. This also means that the end users are aware of the zone of uncertainly from the beginning rather discovering it at the time of crisis.\u00a0The understanding of their lack of knowledge is the golden knowledge all wise models possess and inform to stakeholders. In industry, stress testing is one way to measure how models behave in a state of distress or unusual conditions. However, many times this could be just a mathematical jugglery. The breakdown of financial models during the financial crisis five years ago is a testimony that the models were not wise. They were mostly driven by greed and were not unbiased to begin with. The rating agencies S&P, Fitch and Moody\u2019s were stress testing their models while grossly underestimating the situations to follow. \u00a0They either did not have the knowledge of the reality to follow or were scared to reveal the truth \u2013 as it would have affected their own balance sheets and client relationships.<\/p>\n
Sign-off Note<\/span><\/h4>\nAre we wise now? I wish I had the answer. Nevertheless, to be wise financial services, analysts and the world as a whole can learn a few lessons from Buddha.<\/p>\n","protected":false},"excerpt":{"rendered":"
A Case for Wise Models Lehman Brothers, a financial services company, filed for bankruptcy on 15th September 2008 – exactly five years from today. A company founded in 1850 disappeared overnight. The chain of events to follow brought the overall financial system across the world to its knees \u2013 hence started the biggest economic crisis<\/p>\n
<\/a>Laughing Buddha – by Roopam<\/p><\/div>\n
Traffic can make the polite use cuss words. What is it about traffic that can perturb us so much? Is it that we want to be somewhere and being stuck is just not acceptable? Awards can make the humble proud. Accolades and appreciation have that kind of influence on all of us. These are just a few examples, in daily life; there are tons of things that ruffle us.<\/p>\n
The Buddha<\/span><\/h2>\nWhen I think of the wise, say Buddha, they say he was unperturbed in most situations. You praised him and he stayed calm. You offended him and he was unflustered \u2013 at least that is what most Buddhist stories say. So was he detached and did not care. Is that the secret of being Buddha? He did leave his family to live the life of ascetic. However, then he did realize that it was useless, stopped running and came back to the world. He was absorbing information, as we do, without getting affected. Rather, he provided solace to the perturbed. Possibly, that is wisdom or being wise.<\/p>\n
I think the secret to wisdom or being wise is self-knowledge. Once you know who you are, the external factors would not perturb your inner core. \u00a0It is certainly not about knowing it all but knowing what you do not know. Knowing one\u2019s own limitations and accepting them makes a human wise.<\/p>\n
The State of Equilibrium<\/span><\/h2>\n<\/a>Poise, I have always found this word funny. The sounds of some words tickle your eardrums \u2013 poise is one such word for me \u2013 not the meaning of the word just its sound. Poise actually means equilibrium, state of balance or composure. Let us consider the adjacent graph. The green zone in the middle is the state of poise. The state you find yourself while reading a good book on Sunday afternoon, or playing with the kids, or staring at the night sky. A manic-depressive often lives in the red zones. Most of us traverse through all the zones in a day\u2019s time \u2013 the highs and the lows. Traffic throws us down and awards up.<\/p>\nWise models<\/span><\/h2>\n<\/a>The law of Average<\/p><\/div>\n
Keeping the above in mind, let us come back to analytics. Recall the most fundamental concept of average. Now, consider two different kinds of averages (a.k.a mean) one with million data points and other with just five. Both have the same value of say 100. Now, you throw in some new information and add 1000 to both these values. The first one will stay mostly unaffected but the second one will jump up 150%. The down side or adding -1000 will have the same impact on the values.<\/p>\n
Regression towards the Mean<\/span><\/h2>\nThe following is an example of regression model. The relationship between age of a teenager (in years) and his height (in feet) is displayed in the model. This is most certainly not the best of models but is used for the purpose of illustration.<\/p>\n
<\/p>\n
Now the beta values or 0.1 and 4 in the above case are the soul of regression models. The beta values are actually average or mean values of the parameters. The ideas we discussed above about averages applies to the beta values as well. Notice, this is not a universal model. It will only hold true for teenagers. For a 100-year-old male, the model will not estimate him to be 14 feet tall.\u00a0The same applies to logistic regression discussed in a previous article<\/a>. Genuine models have well defined boundaries for which they will hold true. They don’t pretend to be true universally.<\/p>\nWe have already created the foremost requirement for wise models. They do not fluctuate much when some unusual information comes their way.\u00a0On the contrary, wise models understand the limitations of their knowledge and define the boundary for which they will hold true. This also means that the end users are aware of the zone of uncertainly from the beginning rather discovering it at the time of crisis.\u00a0The understanding of their lack of knowledge is the golden knowledge all wise models possess and inform to stakeholders. In industry, stress testing is one way to measure how models behave in a state of distress or unusual conditions. However, many times this could be just a mathematical jugglery. The breakdown of financial models during the financial crisis five years ago is a testimony that the models were not wise. They were mostly driven by greed and were not unbiased to begin with. The rating agencies S&P, Fitch and Moody\u2019s were stress testing their models while grossly underestimating the situations to follow. \u00a0They either did not have the knowledge of the reality to follow or were scared to reveal the truth \u2013 as it would have affected their own balance sheets and client relationships.<\/p>\n
Sign-off Note<\/span><\/h4>\nAre we wise now? I wish I had the answer. Nevertheless, to be wise financial services, analysts and the world as a whole can learn a few lessons from Buddha.<\/p>\n","protected":false},"excerpt":{"rendered":"
A Case for Wise Models Lehman Brothers, a financial services company, filed for bankruptcy on 15th September 2008 – exactly five years from today. A company founded in 1850 disappeared overnight. The chain of events to follow brought the overall financial system across the world to its knees \u2013 hence started the biggest economic crisis<\/p>\n
The law of Average<\/p><\/div>\n Keeping the above in mind, let us come back to analytics. Recall the most fundamental concept of average. Now, consider two different kinds of averages (a.k.a mean) one with million data points and other with just five. Both have the same value of say 100. Now, you throw in some new information and add 1000 to both these values. The first one will stay mostly unaffected but the second one will jump up 150%. The down side or adding -1000 will have the same impact on the values.<\/p>\n The following is an example of regression model. The relationship between age of a teenager (in years) and his height (in feet) is displayed in the model. This is most certainly not the best of models but is used for the purpose of illustration.<\/p>\n Now the beta values or 0.1 and 4 in the above case are the soul of regression models. The beta values are actually average or mean values of the parameters. The ideas we discussed above about averages applies to the beta values as well. Notice, this is not a universal model. It will only hold true for teenagers. For a 100-year-old male, the model will not estimate him to be 14 feet tall.\u00a0The same applies to logistic regression discussed in a previous article<\/a>. Genuine models have well defined boundaries for which they will hold true. They don’t pretend to be true universally.<\/p>\n We have already created the foremost requirement for wise models. They do not fluctuate much when some unusual information comes their way.\u00a0On the contrary, wise models understand the limitations of their knowledge and define the boundary for which they will hold true. This also means that the end users are aware of the zone of uncertainly from the beginning rather discovering it at the time of crisis.\u00a0The understanding of their lack of knowledge is the golden knowledge all wise models possess and inform to stakeholders. In industry, stress testing is one way to measure how models behave in a state of distress or unusual conditions. However, many times this could be just a mathematical jugglery. The breakdown of financial models during the financial crisis five years ago is a testimony that the models were not wise. They were mostly driven by greed and were not unbiased to begin with. The rating agencies S&P, Fitch and Moody\u2019s were stress testing their models while grossly underestimating the situations to follow. \u00a0They either did not have the knowledge of the reality to follow or were scared to reveal the truth \u2013 as it would have affected their own balance sheets and client relationships.<\/p>\n Are we wise now? I wish I had the answer. Nevertheless, to be wise financial services, analysts and the world as a whole can learn a few lessons from Buddha.<\/p>\n","protected":false},"excerpt":{"rendered":" A Case for Wise Models Lehman Brothers, a financial services company, filed for bankruptcy on 15th September 2008 – exactly five years from today. A company founded in 1850 disappeared overnight. The chain of events to follow brought the overall financial system across the world to its knees \u2013 hence started the biggest economic crisis<\/p>\n<\/a>Poise, I have always found this word funny. The sounds of some words tickle your eardrums \u2013 poise is one such word for me \u2013 not the meaning of the word just its sound. Poise actually means equilibrium, state of balance or composure. Let us consider the adjacent graph. The green zone in the middle is the state of poise. The state you find yourself while reading a good book on Sunday afternoon, or playing with the kids, or staring at the night sky. A manic-depressive often lives in the red zones. Most of us traverse through all the zones in a day\u2019s time \u2013 the highs and the lows. Traffic throws us down and awards up.<\/p>\nWise models<\/span><\/h2>\n
<\/a>Regression towards the Mean<\/span><\/h2>\n
Sign-off Note<\/span><\/h4>\n