Катя

hello welcome to bst live the show for systematic and algorithmic traders glad you could join us today where we're going to be talking about volatility and how we can use volatility to outperform the market and our special guest for today is kyle schultz from algorithmic futures welcome kyle glad you could join us today yep thanks for having me andrew i'm looking forward to the discussion right so where are you coming from today whereabouts in the states are you yeah i'm in los angeles um so i've been out in la for about eight years now uh originally from chicago though and went to business school at ucla so once i was out here hard to go back to uh the harsh winters of chicago yeah i bet i bet so let's dig a little bit more into your background as i mentioned we're going to be talking about volatility and you know as i was speaking to you before this call uh you seem to have a lot of knowledge about volatility so i'm glad that uh we're going to be digging into that a lot deeper today but before we get to that how about uh can you give us a little bit of background on yourself how you got started in trading yeah absolutely so i mentioned you know i'm originally from chicago and chicago has such a vibrant trading scene given the cme and the cboe um and the exchanges there so there's you know probably 50 to 100 prop trading firms and it's just a really good community um my brother's friends worked at different prop training firms and i saw them they were doing very well and uh you know left work at 3 p.m and worked for themselves and i said wow that's uh that seems like the dream job i want to do that and went to undergrad at university of iowa and when i graduated it was 2008 right so it was a time when those prop trading firms they weren't looking for traders anymore they were looking for programmers everything was moving systematic and i just didn't have that skill set um so i ended up taking a job at an investment consulting firm and then i did hedge fund manager research for a very large insurance company so that's that's how i cut my teeth in the industry and i was interviewing large hedge funds like aqr brevin howard and just kind of soaking up all that knowledge and um you know that's where that's that's where i kind of cut my teeth but i always want to be closer to managing the money so i went back to business school to ucla that's what brought me to la and started learning programming from some of my data analytics classes so so i'd start to back test different trading ideas in r and python and then i had a stint in equity research and a stint in private equity i did that for a couple years but couldn't get rid of the trading bug right so you know started my cta ravinia investment management and then i started algorithmic futures in early 2019. right okay so can you give us a little bit more insight into the the styles of trading that you do and the markets that you prefer to trade yeah um so you know my trading journey i've explored so many different styles and strategies right everything from option spreads on equity index futures options on stocks commodity trend following mean reversion pair trading but ultimately now i'm focused on intraday and short-term mean reversion of momentum strategies and equity index futures treasury futures and currency futures um so that's kind of my primary focus right now and that's how i've developed the portfolios algorithmic futures um to create diversified portfolios using these strategies yeah so why did you pick those markets in particular yeah that's a good question i i started off my cta trading relative value strategies and volatility futures so calendar spreads on vix futures iron condor option spreads which is a four-legged option spread and these strategies were good but i knew i needed to beef up the portfolio and add some long vowel exposure to the portfolio and longbow strategies are the holy grail of trading right because you can significantly outperform during periods of market crisis events um so during my research i actually found out that the best long vowel strategies are actually not through trading vix futures it's by trading breakout and intraday momentum strategies and equity index futures um so you know there's that old adage the market takes the escalator up and the elevator down right so we have pretty violent down markets and that's where intraday momentum strategies can really outperform um so it was kind of a long journey in exploration but like ultimately i wanted to create a portfolio that traded that could outperform in market crisis events but also had some mean reverting strategies that would perform out in markets like we're seeing in 2021 intraday momentum strategies you know as we were talking about did very well last year but creating a diversified portfolio so that you can outperform no matter what market environment you're in that's essentially what we're the goal is at algorithmic futures um yeah yeah so was that was that one of the key drivers in in the reason why you include volatility as such a main component in your trading or there were other things there as well that kind of directed you um to that focus yeah so you know i think whether you're trading volatility or not you have to look at your portfolio and your trading strategies and and bucket them either into long vowel or shortfall right because at the end of the day your strategies are either gonna underperform or outperform during market crisis events and you know eighty percent of the time we're in a bull market so i'm not saying you shouldn't have shortfall strategies those are important in the portfolio as well but if you're overexposed to short bowel say you're an option seller um you're going to get crushed in 2008 volleybands cove in in march and april of last year so i look at the portfolio and say okay if i'm trading ten different strategies how many are short vowel how many are long well how do they fit together through correlation analysis um and how do you really construct a robust portfolio by doing that um yeah so you just mentioned 2008 there as an example of what can happen do you have any other um you know insights from history that kind of we can learn from as far as volatility yeah i think there's been first of all you have to categorize volatility events uh each volatility event's unique and different right so in 2008 we're in a you know one to two year bear market same with tech bubble in 2001 to 2002 and then you also have very violent one-day drops like the flash crash or the 87 black monday market crash um and then you have more frequent moderate volley vents which are 10 corrections right so understanding how your strategies perform in each of those volatility events is important um you can't just there's no one-size-fits-all like this is a long vowel strategy and it catches every vowel event right so looking back at history um some people may remember valmageddon that which was february 5th 2018. um and so i don't know if you remember what happened in the market but the vix actually spiked from 18 to 37 and you might say well that's you know vix got up to 90 during 2008 but the the big takeaway from that is that's over 100 spike and over the previous five years from 2013 to 2018 you had all these short volatility sellers piling into xiv and svxy the et the inverse vix etfs which were billions of dollars and essentially you know that was up 500 and then in one day that those uh etfs capitulated uh because vic spiked 100 and you know all these easy money traders that were making a lot of money got wiped out and uh you can look at the the reddit threads you know i think a lot of people had a tough time even even a lot of large managers like there was one cta ljm partners which was a billion dollar cta that was selling options they got wiped out during valmageddon so it's an important lesson it's like that that event didn't happen historically right we never had this extremely extremely low vol environment from qe endless qe um where the vix was trading below 15 you know most of 2017 and then even a small spike can do a lot of damage and that's what we saw so the lesson is know your short vowel exposure and really understand how your strategies perform during every volley event that you can look at yeah yeah okay that's a good example so thanks for sharing that one kyle so i can see in the chat there are people asking already how you apply volatility in your strategies so how about we get a little bit more more practical now um how do you actually apply them in your strategies do you build strategies with a specific environment in mind or do you apply it later or how do you how do you fit volatility into the process of building strategies yeah so there's many ways actually i mean the first is you can use the vix index or the vix term structure as a signal so i have you know strategies that do that but i think more prevalent is using volatility regimes in your trading strategies so if you develop a strategy and you're back testing it and you're looking historically and you see that there's different there's the performance is bifurcated during different market volatility events you might use a volatility filter right and that could be based on the vix index or it can be based on some of the futures if there's too much noise in the vix index right another important example is adapting your take profits and stop losses depending on the volatility environment so you know you might have tighter profits and stops during low value environments but you want to expand those during high vol environments when so you're not constantly getting stopped out the market's swinging more and you really have to kind of expand where the profit targets and stop losses are in order to have a strategy that's right sized for the volatility environment that you're in yeah so if you were to apply volatility as a filter would you do that at the beginning of the process or at the end why why would you do it like that yeah typically it's going to be at the end because i want to look at the back test and understand how the performance of the strategy looks and you know if i can make that inference that i notice different types of performance during different vowel events then you know i can start examining and looking at different um i might look at okay how does this strategy perform if the if the vix is above 40 or below 40 right because you you want to use a number where it there's not too much noise in it or you might adjust your uh profit targeting stops based on you know where the volatility level where the vix index is at um so it can be it can be pretty um it can enhance your strategies uh pretty significantly yeah so what if you do that analysis so you built your strategy and then you you're applying the volatility filter to see how the strategy performs in different regimes what if you identify that there is a particular regime where the strategy is really really poor how would you manage it then because you've got options like you can switch it off or as you mentioned uh just a moment ago adjusting stops and exits what do you do how do you manage that when you've identified a environment where the strategy doesn't work so well yeah i think you can one like you said you could use dynamic position sizing to change to kind of lower the amount of exposure you're trading in a volatility environment where the strategy might not perform well right so if you have intraday momentum strategies that do very well during uh trending and momentous markets and high volatility but they don't do well during choppy um mean reverting markets you can adjust the position sizing or even turn them off right but ultimately if you create a diversified portfolio of you know both mean reverting and momentum strategies you can use some of that but sometimes you just need to diversify away that risk with other strategies right because you're not you you want to avoid over fitting of course and then you can't really um create a strategy that's going to outperform in every single market environment so you'll go i mean you'll you'll overfit your models if you do have something like that and sometimes it's better just to just to leave it alone let it trade reduce exposure but use other types of strategies that are uncorrelated yeah yeah there's a question in the chat here i just want to put it up on the screen from dr amir what type of volatility filters are we talking about here so usually when i'm when i'm talking about volatility futures um i mean volatility filters yeah filters we're essentially looking at categorizing what vix regime we're in right now so you might filter and say okay if the vix is under 20 and we're in a low volatility environment increase position sizing to these um strategies because they tend to perform better in in low ball if vix is between 20 and 40 you're probably in some pretty choppy markets um so maybe some mean reversion strategies that perform well in choppy markets um might might work well and then if you're above 40 you're probably in a in a pretty high vowel environment where momentum strategies might perform well so maybe you'll increase position sizing there so it's really laying on top of filter to categorize how the strategy performs in each of those types of regimes yeah okay we've got a question here from sl this sl style my next question let's put it up on the screen thanks sl other than vix what is your favorite indicator to measure volatility of markets and stocks such as atr state standard deviation or something else so i i use a lot of times vic's term structure like i mentioned vix can have a lot of noise in it so it can create a whipsaw effect but if you use first month second month third month vix futures they move less uh they move less than the vix index right so you're removing some of that noise that can be good i haven't used a ton of atr but i'm kind of currently examining it so it's a great um if you're looking at the range of bars historical bars it can be very advantageous i think who's it jay wells wilder invented this in the 70s right so he had a book called new concepts and technical analysis and you know he's he developed like adx rsi atr so i mean some really good technical indicators that have withstood the the test of time so that's that's a great book to explore that and i think atr can work very well if you're looking at historical bars the difference the difference in my opinion is that volatility since it's meant the vix index is measuring implied volatility of options it's a little more forward-looking if you're using atr it's really historical looking and each can work well so i think atr can be a good tool in the toolbox so to speak yeah okay thank you uh you actually just answered a question from google user wanting to know do you focus on implied volatility or realized volatility for intraday strategies did you have anything to add to that one no i think i think it's just important to note that um the vix index is using implied volatility right so if market makers are sensing that there's a volleyball say an election or any type of market event in the market um that's gonna the fixed term structure is going to increase and the vix the implied volatility is increasing that they're pricing into the options so subsequently the vix index is going to increase so that forward-looking implied volatility analysis i think is pretty important because you know having just looking historically you know you get some hindsight bias and um you might not be taking into account certain events that are coming up in the market so i think vix index really incorporates that yeah yeah okay one more question from the chat before we move on you've already touched on this a little bit but maybe you can clarify a little bit further so this one's from sl again do you filter trades using its own volatility or volatility compared to overall market volatility as well yeah that's that's an interesting question i haven't really used the volatility of the strategy as a filtering i've typically looked more generally at the market volatility as a general measure um but yeah i think i think that would be something very interesting explore i haven't i haven't looked at that and uh it could be could be pretty interesting yeah okay all right well we might move on there's a few more questions in the chat which we'll save for a little bit later so keep them coming got some great questions today so i wanted to ask about i guess a a common approach for getting exposure to volatility was to sort vix futures do you use that approach and what what do you think is the best way to get exposure to volatility so yeah like i mentioned a lot of the short vix future sellers got wiped out in valmagen so you have to understand that that's a short that's a short volatility strategy um you can do it but i think you know instead of just shorting the front month vic futures you can use calendar spreads in vix futures um so you might short the second or third month and go along the fifth month or you can short the back end of the curve which is is moving much less than the front end now the front end you're going to have higher role yield but there's also way more risk in that right so you just you know like any other trading strategy balancing greed and fear in the market and not getting too greedy short vowel can be very good i think there's better ways to trade volatility you might want to look at trading vix versus e-minis right because that spread is about 80 correlated inversely correlated i should say so you can pair trade those um another way is to pair trade vix futures against v-stock's futures v-stock's futures is european volatility um so i think we've got a slide on this maybe you can pull up oh yeah sure so as you can imagine us volatility and european volatility is highly correlated or to use a better term is highly co-integrated meaning that it's a mean reverting spread right because you can have two correlated strategies that don't mean revert so co-integration although related to correlation means that the spread is mean reverting so here on this slide you see vix futures and v-stock's futures and you can tell how correlated and co-integrated the spread is and the spread in the futures is in the green so when you're looking at the spread this is a great statistical arbitrage spread right where you can create say a 20-day moving average of the spread and then every time there's a two standard deviation move trade the mean reversion of that spread so i think the next the next slide shows an example of how a trading strategy could work in that um i think it's important note that v stocks doesn't have a ton of liquidity so this is definitely more of a part of a prop trading or individual trading strategy i don't think a lot of institutions really trade this but it's it's a really unique trade and here you can look at the moving average of the spread and the two standard deviation bars on the upside and downside now if you had a strategy that just shorted when it hit the upper threshold and then closed out when it hit the moving average and vice versa on the downside it could be a really profitable way to trade vic's futures and be somewhat hedged right yeah yeah right um let me just click that off so my apologies there so um you're talking a little bit there about uh when to trade mean reversion and momentum have you found uh in your research that one is is better than another if you if you aren't able to monitor the spreads as you just mentioned there like if you're using maybe a more simplistic approach like atr or something like that have you found mean reversion and momentum work better in high or low volatility regimes yeah so typically going back to that bixby stock spread you have to be able to hold through to when it to when it comes back to the moving average right so if the spread blows out typically in volatility environments high volatility environments the spreads will blow out more so they underperform temporarily while the spreads blow out but as long as your position size correctly you're going to catch that mean reversion and do well but in general mean reversion strategies are relatively short valve because of that blowing out of the spread momentum strategies typically have positive convexity so what i mean by that is that like a with a linear change in the price there's an exponential uh benefit of the strategy p l right so as the market does worse and worse during march of 2020 or april 2020 um you know momentum strategies are are catching those violent falls in the market and you can outperform pretty significantly so i view momentum strategies as essentially long vowel strategies and then other mean reverting strategies as more short vowel or relative value strategies okay let's uh talk a little bit more about um the vix term structure as a regime filter if you've touched on it a little bit already but can you explain a little bit more about how you apply that into your trading yeah um so vic's vick's term structure 80 of the time is in contango um so maybe you can pull up the slide of the term structure just so everyone understands the difference between contango and backwardation so next slide isn't it here we go yeah so here's here's an example of contango where the vix is at 15.57 and you see the vix term structure um is downward sloping which means that as these futures approach expiration you're going to get a positive role yield by shorting the vix future um and then the opposite with backwardation so 80 of the time the market's in contango um usually when there's low volatility and then if you move to the backwardation slide when vic spikes the the term structure moves to backwardation where it's downward sloping where it's upward sloping sorry um and so looking at the term structure you might compare say where the second month vix future is compared to maybe later in the term structure but usually when the term structure's moving into backwardation maybe that's a cell signal right and you don't always want to wait till it fully hits backwardation but as the curve is flattening your signal could be your exit signal could be based on the vic spot compared to the second month future and when the ratio between that um gets closer to one you might exit a strategy ahead of a huge market downturn right um so that's that's one way you can trade it and incorporate this term there's a lot of information in the vix term structure so there's there's so many different ways to look at different maturities and and use that as a signal yeah so just on this slide here so you just mentioned that you might want to compare say the second month there with something further back how far back would you go because as you can see here that it kind of goes flat after a while so does that or does that use so does that lose its um relevance after a period of time yeah the back end of the curve really doesn't move much i mean vixx the vix is one of the only mean reverting asset classes out there which creates this unique um unique um structure of the term structure most commodity term structures don't move from backwardation to contango in a mean reverting fashion right if we know the vic spikes and we're in backwardation we know eventually it's going to come back to contango um so that's very unique aspect of vix futures but to your point the back end of the vix curve is not moving that much the front end of the curve might be moving too much and there might be too much noise so really i try to focus on second third fourth month vix futures because i think that's the right balance where you're there's some movement but there's not too much noise in the in the term structure right so you're looking for a ratio between those three if you're doing second third and fourth or how do you how do you how do you apply that algorithmically yeah exactly like you said i use a ratio because that you can use that ratio to tell you the fixed term structure is flattening and it's time to get out of a strategy or it's coming back into containgo and it's time to get into a strategy so entries and exits can be based on on that ratio and typically it's a pretty strong signal um to use at iphone yeah okay so how would you adjust your entries and exits then if the regime changes yeah so so it's interesting because we're moving into a more you know during call it 2013 to 2018 right we were in an extremely low vowel environment and now more recently there was some you know you had vamageddon in 2018 a lot of volatility in q4 2018 you had covid when vic spiked up to um 82 83 um in march so i think going forward you know as we've seen inflation above five percent and i know there's a baseline effect there measuring it off of the cove at lowe's and the fed's beating the drum on inflation's transitory but look consumers when you go to the pump and you're buying food consumers feel it there's inflation out there and the fed doesn't really focus on it or they just want to continue to pump the economy and make sure that we ride out any bumps in in everything we're dealing with from an environment from cobit perspective but i think there is going to be more volatility going forward and as interest rates uh if inflation is in fact not transitory and we see higher inflation the fed's going to be forced to raise rates and you know with zero interest rates for the past 15 years or however long it was there's been a huge shift from fixing them into equities and now we're going to see a reversion of that and and fund flows really can drive equity market volatility if all that money shifts back you know institutional investors pension funds sovereign wealth funds if they start moving back into higher yielding uh fixed income securities they're selling their equities and those fund flows can drive future volatility so i think going forward we could be an interesting market environment where there is more volatility um and you need to pay attention to some of some of the term structure and the vix index when developing your strategies yeah so the term structure then you think will be the um you know the major sign that that things are changing there or you're looking at other things as well no i think i think the vix term structure is probably the most important right because um when if you're seeing that the term structure is not in steep container when when we're in a very low volume environment there was very steep contango in the term structure right and so there's there's information in that now i think what you're going to see going forward is that you might not have that same role yield and that same steepness in the term structure it might be a little bit flatter it might be moving around a little bit more and also it's important to look at some of the kinks in the term structure because what that really represents is event risk in the future right so if um some of the vix futures at certain maturities are higher than the whole term structure maybe there's an event there that you need to pay attention to so use that information in the term structure in your trading i think can be really helpful okay uh we might just go to some questions in the chat now um so here's one from jeff thanks for the question jeff do you use volatility to trade opening range breakouts such as futures at new york open currency futures at london open dax open etc so i i do trade breakout strategies but i don't use volatility in those strategies particularly i use volatility more in mean reverting strategies and then also asset rotation strategies so if i'm looking at a strategy that might trade e-mini s p 500 futures treasury futures and gold futures you can use that vix term structure to rotate out of equities and into treasuries or gold as a safe haven right if the vix term structure starts to move into backwardation so i found that to be the best for opening range breakouts i really don't use a ton of that i actually going back to adx i think adx is is a great um technical indicator to use in in momentum strategies right um because that's moment that's kind of evaluating uh how much momentum there is in the market so more for mean reversion and asset rotation type strategies and less for momentum strategies yeah okay we've got a question here from neil um you've given us a couple of examples earlier about you know volatility exceeding or coming below particular thresholds but what about regimes that are rising or falling perhaps above or below some moving average yeah so so like i said if you look at historically different volley vents um i found that you know vixx tends to spike very quickly um and then it might come down gradually over time so it's hard to catch that big spike all the time i don't really use long boss strategies in that perspective where vixx is spiking the term structure is moving into backwardation so now i'm going to go long vix because you can really get crushed when vix collapses right um usually those big spikes are very small they're very short and then over time they can they can move back i mean if you look at vic's spiking to close to 85 in march of 2020 it's now back at 16 right so we've had a long you know 18-month environment and there's been some spikes along the way um but it's been gradually declining and it's really hard to catch those spikes i use it more as a signal of when to get out of certain strategies that are short vowel or have some type of beta market exposure right it's it's more about hedging that tail risk getting out of the strategies that don't perform well during during volatility environments okay a question here from hitesh this one was very early on in our discussion have you modeled volatility clusters so i haven't i think i think using some machine learning uh classification could be incredibly interesting for for modeling valve because you can really create some distinct classifications and use that to create strategies for each of those classifications like i mean you've had ernie chan on the podcast he puts out some great content and he talks about how machine learning you know it doesn't always work well in signal generation um you can use random force and classification uh machine learning strategies to create um to create markets and regimes and identify and alter your strategies that way but um those types of that type of research although it's although it seems interesting and there's a lot of advances there i don't think it's really been used in signal generation effectively so you know volatility clustering could be very interesting to separate regimes okay thanks kyle uh a question here from ola thanks for the question ola ola is a regular how do you get historical information about the vix term structure yeah so vixx central those charts that i pulled up i think vicscentral.com is a great source where it it has all of the the vix term structured data out there now i use i trade on tradestation right and i subscribe to to the feeds to get all the futures data another source i think you can use is quando so back when i was back testing strategies in r and python i would call from cuando for most of their vix futures data and incorporate that into my back tests i think they may have changed it it was free but now i think they're charging for it um so you know you can use that i'm not sure how expensive it is but vic central is is a great resource as well so you can go there and it shows the current term structure or you can enter in a date and i'll show you the term structure from that date it'll show you different volatility indexes not just the vix but the medium term i think it's vxm and the short term vix so that's that's a great resource in my opinion okay here's one from google user do you use any dark models with vix to improve its predictability so i i do not i know garch modeling um and that's generalized um auto autoregressive conditional with the heteroscedasticity there hence why hence why we use the term garch it's used a lot in volatility forecasting i rarely am forecasting anything i'm using statistical analysis to understand how strategies behaved historically um and creating a diversified portfolio of strategies while being mindful of not overfitting so how do you not how do you avoid overfitting is a good questions i you know i only use a couple parameters i use walk forward analysis to change the in sample and out of sample period for my back test and then most importantly and probably the most simple method is to change if you change your parameters slightly how much does this change the back test if you're seeing huge swings and changes in the back test it's probably not robust um so i don't use scotch modeling i don't really use any forecasting i'm not trying to forecast volatility but i know when you talk to different quantitative investment management firms that's used pretty heavily as a method i'm really doing statistical analysis historically creating diversified portfolios and um you know looking at looking at the trading strategies from a portfolio theory perspective yep okay thanks kyle here's a question from a trader in brazil brazil ronnie i think it is ronnie irani i'm a trader in brazil how do you manage your risk yeah uh great question um because everyone has a different approach right so i manage risk by looking at historical back test and i'm really targeting a specific drawdown um and i trade futures right so i'm trading on margin and you know you can use you have enough leverage you know to get in a lot of trouble with futures so i'm creating strategies combining them in a portfolio and then trying to understand the drawdowns and the correlation of those strategies now it's a great story to say hey we've got a bunch of uncorrelated strategies we're going to put them in portfolio and everything's going to work out nicely right and two strategies might have a zero correlation over the 10-year period you measure great but if those strategies have a correlation of one during 2008 guess what diversification is not protecting you when you need it most um so you can't just look at long-term correlations i like to look at rolling correlations um and then really i just like to pick all the drawdowns and the volatility events and understand how the strategies performed really examine the trades when did it trade during the day when did it hit the stop loss and i look at i look at daily data to understand the left tail in my back test right so what are the worst days in the back test and why what happened on that day so it's a less although i'm a quantitative and systematic trader it's a less uh quantitative uh approach and a little bit more subjective of understanding your strategies intimately yeah yeah you raise a good point actually about correlation uh measurement because you know i see uh lots of traders that will measure the correlation their strategies over i don't know five or ten years and as you mentioned it can be very low number um but uh but it's a rolling number right depending on the the regime so how do you actually um how do you how do you do how do you measure that like do you actually uh split out the regimes and then measure the correlations or do you have like a a maximum minimum or what do you do to analyze the correlations if you're looking at a couple of strategies yeah typically i'm just you know looking at the correlations on a rolling basis so i'll put the returns even just simple excel work right put the returns in excel look at the correlations and how they roll over time and you know if if they move up significantly what type of market environment is that in and do you want to trade these strategies together in a portfolio if it's output if it's underperforming um in a high vol environment right so you know everyone uses correlation tables a lot but you can you can really um kind of fool yourself by using those and and just the law of averages right don't always hold up you you need other statistical analysis so rolling rolling correlations is one of them at algorithmic futures um recently i just did a correlation analysis for a client of mine he's trading a bunch of different systems and he wanted to understand how algorithmic futures performs relative to his other systems so so i think he uh he got a lot of insight from looking at the rolling correlations um and seeing how the strategies perform yeah okay thanks for uh clarifying that so we've got a question here from mars explorer love the name um what is the best indicator to measure volatility if a stock a stock except for vix yeah i mean standard deviation is the standard measure um but it's not the best right because if you look at if you look at atr you're really average true range you're really looking at the historical bars and the range of those bars so i think you can get a lot of good um robust signal out of that to understand um on a daily you know over the past 14 days what is what is the average range of the bar bin um and how is that going to affect your trading strategy right are the bars really tight are we you know either open and close very far away um or the high and the low very far away but we're closing at the middle of it so you know you have to be look at bars from a path dependent perspective not just after look where the closes and the open is is like how did it move up and how did it move down during that bar and i think average true range can be can be very helpful okay and then one one one other thing i'll mention is that when you're when you do an analysis of portfolios a lot of people like to use standard deviation and sharp ratio which is uh returned to standard deviation um but if you've got if you've got a sharp ratio that's being hurt by upside deviation you know that doesn't make sense you you i'll take upside deviation any day positive returns high high positive returns sound good to me right so you want to use sortino ratio which uses return to downside deviation or kalmar ratio omega ratio that's looking at the distribution of returns there's a lot better metrics than sharp ratio although sharp ratio is kind of like the industry standard for allocators and uh you know it's uh it's very misleading yep yeah absolutely all right we got another question here from jeff here we go coming up when do you decide to quit trading a specific strategy so i i guess i would i would kind of i'll answer that in two ways if you're talking about when do i exit out of a strategy is i trade always with take profits and stop losses right i create very structured trades and going back to how do you manage risk that's one way when do i decide to quit trading a specific strategy for example if a strategy isn't working um you know usually i really haven't had any issues with i do so much upfront research that when i create a portfolio it's very rarely that i stop trading a strategy i understand the periods it's not going to perform well in and i think that's the best way to get comfort in in your strategies right if you're just if you're just switching up your strategies or re-optimizing them every time you have a drawdown it's kind of like playing a playing whack-a-mole with every drawdown right that's not a way to develop robust portfolios in my opinion um but occasionally look it happens right i mean if if markets change significantly or regime changes that you didn't anticipate you might just need to go back to the drawing board examine the strategy and be truthful with yourself and say hey i shouldn't be trading this anymore it's not trading well in this market regime you know trading you're you're graded every day uh with with trading right so you got to be truthful with yourself and and uh really examine your strategies yeah so a couple minutes ago uh we were talking about the um the the current state of the environment and what could be happening with all the other you know all these things going on outside of the markets uh so what do you think traders should be doing to get prepared for you know anything that could possibly be coming in the in the future so i think you know history history rhymes it doesn't it doesn't repeat but it rhymes right so you want to look back at periods where we did have increasing inflation or where rates increased if you're if you're back testing over the past from like 2008 to current um you're really picking a specific market regime you're back testing your strategies on and you might not be prepared for what's coming ahead so if you look back to the 70s um you know that's a period where you had very high inflation and the fed increased interest rates significantly and what you had was i think in 70 and 71 the market dropped 20 percent and then in 73.74 you had the oil embargo that caused significant inflation market was down 40 um and by the time you got to the end of the the decade of the 1970s real returns on equities were negative right so here is a high inflation rising rate environment um and it could provide a glimpse into what we might see in the future i'm not saying that's what we're going to see we don't even know if if inflation is going to be persistent and so you know the fed obviously obviously doesn't so it's helpful to look historically to see what type of volatility you might see in the market if inflation's increasing and rates are increasing because i'll tell you what it's not going to look like the past 12 years or even from 2009 to 2018. right so from a practical algorithmic point of view how do you test that because i imagine data in the 1970s is pretty sparse or accurate data anyway so what can what can algorithmic traders do to get prepared to prepare their strategies you can use different stress testing um i think if you have the data a lot of times when i'm using tradestation data you know equity index futures go back to 97.98 so you're getting a longer testing period right so i think it's more about back testing throughout longer periods just to see how strategies will perform and behave i'm not saying you want to optimize during that whole period but it's more about examining it and and looking at that so other other areas you can use monte carlo simulation stress testing in the portfolio um and those are also ways to look at the risk of your portfolio but from an algorithmic trading perspective back test and look at your back test during um i think we had you know a little bit inflation 2004 to 2006 right so there's periods that if you have a longer um back test period it might provide some more insights okay all right we'll ask one more question here from the chat and then we'll uh start wrapping up from for today so here's a question from ola thanks ola what is your method of allocating funds between a set of strategies yeah that's a great question um and there's a lot of debate about this right because essentially if you're dynamically dynamically trading strategies there's high probability you might be overfitting um so i think it's good to be cognizant of that i personally you know as i create a portfolio i'm trading futures right so if i look at algorithmic futures portfolios i might have a 50k portfolio that's trading 13 different strategies some of those are intraday some of those are longer term but i'm really trading only one contract per strategy so i can't really if you're trading smaller account sizes you can't get too complex with allocating or increasing contract sizes for strategies and i'm also fine with that because sometimes a simpler approach of equal weighting strategies is a lot of times better right it's hard to predict when your strategies are going to outperform and underperform so you can be completely fine just trading one lots on each strategy and keep it equal weighted if you're gonna look at optimizing between strategies or using dynamic position sizing be cognizant of overfitting because now you're creating multiple layers of overfitting right so you can overfit your back test and strategy overfit the portfolio there's hidden risks to that so a lot of times keeping it simple is best yep okay all right well uh i did say that was the last question but there's one more here i want to ask you and then we will finish up i promise so let's put this one up here for sl um with with trend following strategies and the market is trending up do you prefer to trade when stock volatility is decreasing or increasing do you care if overall market volatility is high or low at that time so if you're trading a trend filing strategy and the market is trending up it is most likely that volatility is decreasing right because you might have had a volleyball now it's trending up and typically like i said um you have violent down markets but very low vol up markets um so i think you definitely want to see that um and that could be a good confirmation signal is okay vix is declining we've got a good trending market maybe you're using market moving average crossovers maybe you're looking at price action but having a confirmation signal of volatility decreasing is is probably a good um enhancement to the strategy okay all right well uh thanks for your time today kyle how can people uh either get in touch with you or learn more from you where can they go yeah so people can check out my website which is www.algorithmic.futures.com and just note that there is a dash in there and here's you know a lot of information on the trading systems that we offer our clients and you can also email me at info algorithmic dash futures dot com once again there's a dash in between algorithmic and futures so i urge people please reach out you have questions you want to learn about the trading systems that we're offering um i'm glad to chat and and uh always uh always open the conversation yeah okay great thanks for that because i can see in the chat there's a couple more questions that we uh didn't get uh time to address from so for example from good old boys so uh they can contact you via email or through the website and uh it's good to hear that you're happy to answer more questions so um yeah thanks very much for your time today kyle really enjoyed the discussion uh was there anything else you wanted to add before we finish up for today no i i look i had a lot of fun and uh this was a great discussion so i appreciate having me on the podcast and um you know i look forward to any any um listeners that do want to reach out and chat more okay excellent we got a lot of good feedback in the chat so let's have a look at a couple of uh thank you before we uh before we finish up so here's one from sl thanks kyle thanks andrew jeff said uh great conversation thanks for the questions today jeff here's a good one from ola thanks for sharing your wealth of knowledge and experience and thanks for another great session andrew and um good boy have some questions for you i'll email you but thanks so much for your time and final one this one's from neil nil had some good questions thank you nice and simple so yeah um yeah it was great chatting to you uh today kyle and uh appreciate the um the knowledge that you've shared with us and i wish you all the best yeah thanks so much andrew take care all right cheers thank you bye you