 Okay, all right. Welcome to the final SE3X live event. Thank you guys. I'm sorry that we're running a little bit behind. We have, we had a few technical difficulties, but it looks like we are all set to go. So just a few more reminders before we dive right in to our, to our event. As you're all aware, the final exam has just been released. So you have exactly one week to finish it. It'll close on September 27th at 1500 UTC as normal. So make sure to do it before then and you only have the allotted four hours. So once you begin it, you have four hours and you know, there's no extensions for that. So make sure you, you know, block at the according amount of time. And so just remember, if you're starting it two or three hours before it's due, it will close at 1500 UTC. The other notices are that, you know, week 10, I had an email saying, we, you guys struggled a little bit with it. So we've actually gone in and given partial credit for many of the multi select questions. So that's a bit up. So the average is about 82% now. So if you guys go back in and take a look, the answers are available. And also we edited that a little bit. So now let's get right into the topic of hand, because that's far more interesting. So, you know, today we're very lucky to have Dr. Jose Velasquez Martinez. He's, I'm sure as many of you guys know, the director of the supply chain management blended program here at MIT. Of course, just one quick note, you know, I know a lot of you guys might be pressing ask him questions. Unfortunately, today, the topic is not on the blended program. So let's keep our questions on hold until there's a next event where you can ask him about the blended program. Today we're going to be talking about the screen game that was available in week 11. And many of you guys filled out the survey on what you found. So I'm going to kind of cut it short and give it right over to host way so he can introduce the game and some of the debrief. Okay, well, excellent. Thank you, Alexis. It's really a great pleasure to be here with you guys. And I really hope that you enjoy playing with the game. So part of the idea is to complement the teachings on supply chain risk management. Let me just get rid of my earplug so that I cannot hear myself. Anyway, so yeah, so this is a compliment, as I said before, and just before starting the discussion, I would like just to make an acknowledgement right on the situation that is happening now in, well, in my home country, Mexico. As you know, we've been hit by a couple of earthquakes and this second one, despite the fact it was not as hard as the first one, it's still the situation is being really difficult because it was extremely close to the main city, the Mexico City. And well, you know, it's always very hard to predict as I prepare already some of these discussions with you in supply chain risk management. But still, I wish well to my home country, a country man. And I hope that if some of you would like to help, you can always go and donate and also just to send your prayers. So anyway, so I didn't want to start discussing this without making this proper acknowledgement. Now that we are discussing about this situation, some of you were wondering, and I don't receive some of your comments and questions via email and also in the server that you fill in, that it was very hard to really come up with an interesting solution for the screen game without having any information about the type of scenarios that we were planning to use to test your policies. And this is true. And one of the main points why we didn't want to do it is because it's part of the training of this game to teach you or to give you the experience of being in a situation in which you don't have any other information, but just your common sense. And that's it. Right. And if we ask my country man from Mexico, how difficult is to actually forecast the impact of an earthquake and also when this is going to happen is really hard. It was just a huge coincidence that actually happened September 19. And I think it was 22 years ago when in 1985, I'm not mistaken, we were hit exactly, but an earthquake exactly the same day that destroyed completely half of Mexico City. And actually the damage was very huge. But the interesting part is that you can always try to build correlations, right? And to say, well, by the way, September 19 is the birthday of my sister. So two times in a row, when it was her birthday, we got hit by an earthquake. So what are we going to avoid to stop the earthquake? Let's stop the celebration of the birthday of my sister. So remember that correlation doesn't mean the relation. This is an interesting insight. So in order just to show you what would be like the difficulties, not just of not having this information at hand, but imagine that I will give you some information beforehand. So this is part of what I would like to introduce you with this slide. So I'm going to give you some examples, asking specific questions, 10 questions. And I'm going to ask you to provide 90% confidence interval on your answers. Okay, so I'm asking I'm giving you the 90%. And I want you to give me the answer now that you know the probability, right? So just think I don't want you to go and look in internet for the answers. It's just mainly for you to try to understand and to come up with an interesting answer. Okay, we'll share now this in the in slides. So in your 90% confidence interval, tell me what is the what was the age of death of Martin Luther King? The length of Nile River, number of countries that are members of the opaque, which is this country with the petroleum, number of books in the Old Testament, diameter of the moon in kilometers, the weight of an empty point, an airplane, 747, the year in which Wolf von Amadeus Mozart was born, for those that are in Germany, I'm sorry for my pronunciation, his station period of foundation, L5, air distance from Volota to Amsterdam, and the deepest known point in the oceans. So all these questions were provided by my very good friend, Andrews Nuuk, who is also a lecturer on this screen again, he's a PhD student here at MIT. But try to come up with an idea, provide me with the with the interval, lower and upper bound with 90% of confidence. So I will not ask you to interact on this now, but you can actually share some of your ideas. But if you think about this, what happens is that usually we tend to have a bias in the way that we create our estimations. And at the same time, an overestimation in our capabilities to do estimation. So if I ask you 90% of confidence interval, how many of you actually thought in saying, well, 90% of confidence interval, I will say from zero to 100. But from zero to 100 years, it's actually an interval that has 90% of confidence. But the thing is that in our mind, we say, well, should be between 40 and 70 years, which is not 90% of confidence. So if I show you the results, I can share with you that the age of the age of Martin Luther King's which he was there is 39 years. He was 39 years old. If we go now with another river in Kilometer, where is the length? It's close to 7,000. But then again, all these answers, you know, how many of you actually got, let's say, nine out of 10 of these answers? Or eight out of 10? Or seven? Last time that we did this exercise with very smart people as you are, you know, we did it in the digital program, you know, the Latin American program for students that are enrolled in master programs and they come to MIT for three weeks. And when we did this exercise in 2000, that approximately only one person first got seven out of 10. One person. And then two persons got six out of 10. And then another two, five out of 10. So imagine like five out of 25 actually got more than 50%. And if we make this assessment, it's like, why this is happening? Well, because it's very challenging still to use the intuition of taking this information and to use it to make decision, we usually tend to be overconfident. So in some sense, one of the main points here is, and then we were discussing, you know, we say, well, don't ask people to estimate these probabilities because they don't know, right? And if I ask my countrymen, please give me the probabilities and an echo, it can actually hit again. We don't know it, right? Because there are tons of variables, I mean, there are a lot of potential causes of these that is very hard to predict. But even if I give you this information, still people with these probabilities don't give the right answer. And this is part of the challenge when we are working on disruptions, right? And in this case, disruptions that are in the domain of natural disaster. So it's very hard to predict. And once this is happening, it still is very hard to predict not just the causes but also the effects of these disruptions, right? So anyway, this is just to give a funny and entertaining exercise for you to try to come up with interesting ideas in your mind and to try to understand that it's not easy to keep these probabilities. And even though we may get some estimation, it still is very hard to come up with the right answer. All right, so now we'll come back to the screen game. And as you may recall, the screen game, well, the main objective is to design a robust mitigation strategy. So when we say robust, we mean those strategies that can cope with the majority of potential scenarios and situations, right? Because we don't know what's going to happen until we have an uncertain future. So we need to decide this mitigation strategy with the aim of minimizing the cost at the same time that we are maximizing the set of levels. So this is the main challenge. And you were giving the basic supply chain, you were giving a plant, a distribution center, and also a supplier. And you were giving a supply chain that is already in place with a policy that guarantees 99% of the rate on their normal volatility. And we discussed this with some of you were wondering about this, like, well, you said it was going to be 99% and then later you said something about 90% of the rate, but 99% is already in place, you have already assessed this thought. It worked in process and it finishes. But then we are going to, you're going to hit, right? We're going to have some disruptions in the supply chain. So you need to decide now backup plants, you know, and backup inventories that you are going to have also in the distribution center, also in the plants as well. So you were giving a survey with five digit code option to fill in and the ideas that you will provide exactly what's going to be the amount of backup inventories you're going to have in the plant and in the distribution center and what type of configuration for backup you're going to have for the supply chain. And you were giving seven options, I remember. First option is not to do anything, right? So you are just keeping your safety stock policy and you will not spend any money to try to protect for adoption that may or not occur. And then you were giving others, like second option is I will have 50% of the capacity either for distribution center plant or supplier. And I will have a time response time of four weeks. So that means that once you are affected by a disruption, it takes four weeks before you can actually use these capacities that you have in the supply chain. And then again, you're moving towards the seventh option in which you have actually 1% of capacity, completely available in the first week. And of course, it's way much more expensive than the rest. So how you can actually balance all of this and the idea is that while you are deciding one of these backup options for your facilities, for those weeks in which you are not getting any option to have protection in case of a disruption, then you move for the option of having some backup inventory. So how to mix and how to move on. So we're giving this option, right? So I'm just repeating the same instructions you were giving in the video. So I hope it was clear enough. Okay. So now let's move up with the analysis of the results of the screen game. So these were the participants. So this is kind of for the participants. I will show you all of them. So hopefully you can find your edX ID name. And also, you know, the ID that I'm just writing there is just a number because we will not be able to track all your solutions and we will show you some of the charts and also your solutions compared with the right of front. So it's going to be up to you to try to track your own number. So you can see different type of solutions in terms of the amount of inventory and also the backup facilities you are deciding. So for instance, we can observe in the case of ID number 10, there is a policy that is quite conservative in terms of the inventory, right? Having 1700 units of inventory and then selecting, like for distribution center and plant, let's say an average backup option, like number three and three. And for the suppliers, you are actually or in this case, the ID 10 is speaking the suppliers option six. So it's like saying, I'm way more concerned for what could happen in the disruption in my supplier than actually in the rest of the supply chain or at least in the rest I can use the other option to balance with inventory. There are other options that are maybe more balanced, like for instance, the case of ID 17, Christoph, in this case, you know, the difference is substantial compared with the Ismael in the ID number 10, which is to have like maybe one or two weeks of inventory in finished goods and working process, but having a quite solid option in backup facilities, six, six and six for distribution center plant and supplier. And other options as well, you know, more in between as well, the option of melody. Anyway, so let me just show you the rest. Anyway, so now I have in the screen the participant right. Okay, so you can actually see yourself and see also the policy. So try to keep in your mind the ID number so that you can track the results that we're going to discuss and how your policy is deep under the different scenarios that we tested. All right, so let me just remind you again, how do we assess the mitigation strategies. So this is going to be an interesting information. So now I'm going to describe the scenarios that we use to test your policies. I'm giving this information. You're going to be asked to go to a breakout room and also discuss with your team what would be like the best policy, right? So you will not run again all of this and mainly you will discuss your policies that you already selected and submitted and you're going to pick one of this. I will explain this in a moment. Or a scenario. So these were the main baseline scenarios that we use to test your policy. So in general, you can see here 10 different options and for each of the options we have a disruption that may occur in the distribution center in the plant and in the supplier. So for instance, when we take the first one in the first option, we actually have a disruption that is not happening anywhere. So you say we're starting with one, has a duration of zero and is ready to be online in the first week. So nothing happens for the distribution center in the plant and supplier. But for instance, we take scenario two. In scenario two, we have a disruption that starts in week one in the distribution center and lasts 12 weeks and then the distribution center is ready in week 13. So now if you go to the plant, same thing, right? So in these cases like a disruption that occurs after the disruption in distribution center has finished. So in this case, you have that the disruption in the plant starts in week 14, lasts 12 weeks and then you again, the plant is ready to be available in week 26. And then again, you go like this in the supplier. So you have different scenarios. For instance, it is one that looks quite strong is scenario seven. In this scenario seven, you have a disruption that occurs in weeks 27. But it's actually happening in distribution center in the plant and in the supplier. And then in order, like for instance, scenario nine, you have a disruption that occurs in week one, but then lasts the whole year for the distribution center. So again, let me just tell you why we picked these options. And we did it just because we want it. It is now actually like clear reason. We have for sure examples of situations like these that have happened before in the in the past, you know, with tsunamis, with earthquakes and, you know, with natural disasters and also politics as well and situations like trade agreements, etc. So we have information and this information can help us to create these types of scenarios. But in general, the idea at least in this game was just to create these scenarios and to try just to come up with something that may happen, like you are feeling in a situation that now my countrymen in Mexico are failing. This is happening. When you prepare for this, when you have like all your goods or your, your the facilities that should locate your transportation, like all the configuration of your supply chain. Well, in this case, it's a very basic setting of supply chain. And we are going to just test your policies against this. But let me tell you how we did it once we got this information of the potential baseline scenarios. We actually run again an assimilation for the scenario. So let me just go back again to the slides. And then in this case, for instance, we created something that we call, for example, a sunny day. Sunny day means you will have 100% of probability that scenario one will occur. I'm not there for 0% for the rest. So that means nothing is going to happen. So it's going to be a sunny day. Fantastic. For instance, now we have another scenario. It could be even probability. This considers that all these 10 scenarios would have each of them 10% of probability for humans, which means everything could happen. Right. And then also we, you know, set different scenarios between them. Right. And I will show it to you. So we create a random scenario. We run this like 50 times and then we see how many times actually your policy performed. Correct. So some of you were also asking, you want to set the minimum or the maximum? In this case, we are taking the average of your performance. Right. Okay. So now that you have all this information, then let me show you the rest. You have the order, like partly sunny. That means you will start having 2% of probability that any of the scenarios after two, three, four, five, six until 10 may happen. 2%. A slightly sunny will be 5%. You know, and slightly cloudy is 10%. And then until we go to orders that are like the short order lapping, which is the one that I just mentioned in scenario seven. This again, it means disruptions happening exactly in the three entities of the supply chain at the same time. And so on and so forth. So we use this information. Right. And for sure, many of you also were wondering, it's possible just to decide one single solution. And that's it. Well, it's not a deterministic setting. So we know that in this case, we have an uncertain future and the DS to see, well, I will just create this and using this information, let's test if your policy can cope with this and how many guys. All right. Okay. So let me just again remind you how we conducted this assessment of mitigation strategies. And as I said before, we can always take the solution that is the best in terms of cost and we can take the best in terms of service level. And we can just try to map, right, all the solutions by identifying these lines. And some of you were wondering, what is the field rate that we need to target? And I said 90%. However, I need to confirm that at the end, we change it for 95%. And the main reason is because most of you actually got very good policy. So congratulations for that. I will show you the results in a moment. But really, all of you did great, great work. So anyway, so you define all these areas, remember? So we said if you are in these red areas, it means that, well, either you are still below what I'm expecting, like in the case that you are trying to save money, but you are not actually getting sufficient amount of field rate. Or you are actually not concerned about spending money. And you are still achieving good service level, but still at the expense of high cost. And of course, we also identify this area, which is having like high cost and very low sales level, which I said also in the video, you are in this area, maybe better looking for another job. That's again, just a joke. But it means that you are actually spending too much money to actually the amount of service level that you make it. And then again, if you are now in the blue area, in the blue area, good solutions for a reasonable amount of cost, you are getting a reasonable amount of service level. And then again, in the green area, these are like the very smart solutions, you are getting very close to the price. And then let me just clarify again for adults that are in love with math and all these spaces, solution spaces. This is for sure not a complex area. So I'm just showing this because it's easier to explain this in a slide like this. But you will see also the solutions of your policies. And you can actually identify that the Pareto front is actually a non-convex function. And you will see it in a moment. So all those solutions that are in the green area get two points. Solutions that are in the blue area get one point. And those that are in the red areas do not get any point. Okay. And then at the end, we just score these and we see how many times you are actually getting good solutions for different scenarios. And then we will rank the best and the best we'll get, you know, all respect, prestige. And Arthur is going to fund you a full scholarship to study the axis. But anyway, so now you will get all the respect, right? And admiration. You are actually getting good results. And at the same time, we also build theoretical Pareto front. And then in this case, assuming some deterministic values, we got like the best of all the potential solutions we make it. And we also compare your solutions with this theoretical Pareto front. And I will show you also the results of this. This is just to actually, we decided to use this for this exercise because we got a lot of very good results. So you will see it. And then we needed to find a way to do the hybrid. And this is how we did it. So we'll explain this in a moment. All right. So now taking into account the following questions. So this is some of the things that we want you to consider when you have decided a mitigation strategy. And this is, I think, a good compliment for all the supply chain risk management course that you've been taking in this ACVX. And in this case, think about your policy and wonder yourself if to build this is better just to have safety stock and not to do anything. Right. And well, by this time, I expect that none of you actually is considering option A is still is the one that is the most used option in practice, which means I'm just going for the reduction of cost. And I really do not care too much about the rest because again, I'm solving problems that maybe are not going to happen. Right. And this is always very hard to assess. Then you can always combine what if now I asked my policy back up inventory? What if I add now backup facility? Or what if I combine the option of backup inventory with backup facility? Right. Okay. Other questions. So we saw some of these policies. I just mentioned the first one in the option was ID number 10, if I'm not mistaken. It was a lot of protection, you know, in terms of backup supplier and plants, but very little for distribution center, right? Or maybe more more in distribution center, the backup inventory. So questions will be like, what is actually the most important in the among all these three options in the in the supply chain to be more protected against a disruption in the in the supplier or in the plant or in distribution center? What will be the least important? So take these questions and also consider again all the other policies that you already submitted and discuss this in your backup. I go with your team and try to present your options, discuss, try to explain your reasoning line. And at the end, what we expected that you will come up with answers to those questions and also to provide a single policy. So you can I know that now we are giving you the information of the design and maybe with this information you want to go back to your screen spreadsheet and try to test all this information. Now you can do that after the live event is gone. But for now just go and assume that these policies are actually part of the team, you know, like like the supply chain risk management policy in a team in a company, you're coming up together, you're going to present your options that you've been working on for the last, I don't know, seven, 10 days. And now you're going to decide which policies the one that you're going to invest and the money you're going to put to be protected. So pick one and we will keep track of that option for the next part of the live event in which we're going to present the charts. All right. So I think they did it on time, right, Patu? Yeah, anything to add, Alexis? Okay. Yeah. Now let's just have everyone go into a breakout room. We have three community TAs with us today, Jonathan, Perram, and Ahmed. If you guys can all go in different rooms and just do some help facilitating, we'd really appreciate it. We appreciate you guys joining today. So let's go into the breakout rooms for 15 minutes or a little bit less and then we'll reconvene and have 15 minutes of question and answer with Hostway and a quick wrap up. So hope to see you back in 15 minutes.