Honestly? I'll go ahead and spill it. I can't get into exact mechanics. But I can discuss what almost EVERY successful firm (not just ours) out there does. Where the majority of the edge comes from, with most Quantitative approaches. It has ZERO to do with prediction. It might look that way, from the outside? But it has NOTHING to do with prediction, and everything to do with the Risk Overlay. Well, when we discuss risk overlay (which is where we derive almost all of our edge) ... I mean all of the various decisions that went into the construction of the portfolio as to mitigate risk by ... 1) Why this market trades against that market should lead to positive expectancy in and of itself WITH NO trade put on, just the nature of those two markets together (and thus mitigate risk) 2) The math advantage ... or ... when we win, we win 1.83x our usual loss, and thus, our edge partly comes from a math advantage on Program ABC that is running (and thus mitigate risk) 3) An inverse risk profile, ... meaning we win a LOT on Program XYZ ... but when we lose, we usually get wholloped 3x our loss, and we have a way to defend that loss as well (and thus mitigate risk) 4) When we put the position on in Program CDE (just making up letters to protect the innocent, ) our position sizing for the whole position, as a directional position, means the thing could go to ZERO, and so no stops are used, as we're dealing with the Notional of the entire Asset ... and it would make for a REALLY bad day as far as that individual program? But it's easily survivable. Then the other programs mitigate that loss as well. So what I'm saying, is we trade that asset in that program, with a position sizing that more moves with the way that market as a whole moves, rather than having a stop loss on it (and thus mitigate risk) 5) The "risk profile" of various programs play off each other. In other words, Program ABC may have an advantaged risk profile, and Program XYZ has an inverse profile, but when we looked at the history, and the way the Strategy is designed, while one has a series of losers, is EXACTLY when the opposite program SHOULD be winning, due to the markets being traded. (and thus mitigate risk) 6) We made sure the different programs had different periodicities. Actually, before we even get to that? We made sure there WERE multiple strategies. There is a benefit here, not of diversification of assets. But a non-correlation / diversification of STRATEGIES. This is why you see so many successful firms that are public? Say that they are MULTI-STRAT or Multiple Strategies. Running strategies that are TRULY non-correlated to one another. The strategies themselves play off one another. It smoothes out the entire return series when put together. Each individual program may be "ok", and could trade on it's own, albeit be a bit clunky ... but when put together, there is a natural smoothing that takes place. (and thus mitigate risk). 7) And then ... yes, we made sure that the different strategies run in different periodicites. Program ABC runs on a 30 day program. Program CDE is a 90 day program. Program XYZ is a 7 day periodicity of average trade. Then there is another program, that has a periodicity of YEARS. Another trade, we might be in for 3 days. Again, non-correlation is a HUGE edge, and often over-looked by new traders. We're non-correlated as to strategy AND periodicity of those strategies (and thus mitigate risk). 8) The weight of each Program, within the portfolio; for the reasons of it's individual volatility over time. So in other words, Program ABC has almost no volatility, a Sharpe of 1.5 since 2008, but it only annualizes 6.9%, so there's not a lot of 'juice' there, but it does make for a great "anchor" or "foundation" process (Stupid simple too, it takes about 3 minutes per month to implement ... simplicity is key). Where as Program ABC annualizes 25.9% ... but has much deeper drawdowns. So we let that one run, but we give it a different percentage weighting within the portfolio. (and thus mitigate risk). The strategies themselves? Are very, very simple. I could teach my 19 year old Nephew to trade the mechanics of them inside two months. But there is a LOT going on under the hood as to WHY they work well together, all to mitigate risk. But as what is going to happen in the future? We don't have a clue. And neither does anyone else. We don't know. And we don't have to know. Like a casino (who has no idea what the next turn of any card at any of their tables will produce), we just run various "games" or Strategies, and we know the math on each.
Think of it this way. Hopefully this will help you out. At least I hope it will. Think of owning a Casino. Let's say, for the sake of example? They have 10 tables, 2 shows, and a restaurant. Small Casino, but hey ... let's just run with that. They have Blackjack A Limit Poker Table Baccarat A High Stakes No Limit Poker Table Slot Machines Roulette Craps Keno A Lottery game Bingo Then they have Two Shows. Their restaurant, is a Gordon Ramsey Restaurant. Right? Now is that Casino Gambling? Of course they are. They gamble every very single day. But they know the math on each of their games, right? They know, every once in a while, someone is going to take them BIG at say, Blackjack. Right? I mean, they know that over the long haul they'll make money, because they know the math, and they have set the payout odds on Blackjack, and instituted other rules. But every once in a while, they are going to be taken. BIG. Think of that blackjack table? As a strategy in a portfolio. Then let's look at the two poker games. They can't lose there. They don't make a lot beyond the rake? But that money is consistent, and steady. Their downside, is if the casino isn't full, and people aren't playing poker. But when people play? They make off that rake. Think of those poker tables? As two similar strategies ... or "Programs" within the Portfolio. As a matter of fact, think of ALL of those games, and restaurants, and shows as "Strategies" within a Portfolio. The Casino has NO IDEA what is coming next. If people will go to their shows. If there will be an economic downturn. If someone is going to take them big at blackjack. If people are going to go into Gordon Ramsey's restaurant. All they know, is that when people DO come, and if they play long enough, and enough people walk through the door? The math they have worked out tells them that they will win. The above post where I discussed Multi-Strat portfolio construction? It's similar. I have no idea what is coming next, in my different "Tables" and "Games" and "Shows" and "Restaurants". However, I do know historically, what the math says each program should do. It's accuracy rate. It's payout. What market's they are exposed to, etc. But I want them different. I don't want to open a Casino, that ONLY has Poker tables. I might make SOME money? But I won't get enough of a draw of enough people, to make money. I don't want to open a Casino, that ONLY has Blackjack tables. I might make SOME money? But every once in a while, I'm going to get hit BIG by someone that takes the house, and again, if I only have ONE game like blackjack, I won't get enough of a draw of enough people, to make money. But if I run all of the above together at the same time as the Casino? Meh, I might lose big on Blackjack on one particular night? But at the same time, I made out HUGE because there was a huge rake going on at a high stakes No Limit Poker Game upstairs, and I also made out at the Roulette tables. Or NO one is playing Poker, and there is no revenue there. The restaurant is empty, but thankfully, some sucker came in and unloaded a WAD on the Blackjack table, and lost it all, making for a positive cash flow night. I have different games running simultaneously, and if one loses? Others should make up for it, making me consistent money. Similarly? I don't want to run a portfolio, with only ONE strategy. I might make some money? But every once in a while, I'm going to get hit with that particular programs / strategies maximum Drawdown, and it's going to hurt. If I run VARIOUS games, or strategies? They offset each other, leading to a more smooth return graph. Take an example: I have a strat that is long Gold right now, and has been long Gold since July. I have another program, that is long and short several markets, and this is viewed as ONE position. Do I know what is going to happen with Gold? Nope. Not a clue. The Quantitative history I have says to be long Gold, and in the way I do it? I make money. But will THIS month be a month that it makes money? I have no idea. And if it turns out to lose money this month? Then maybe that multi-market strategy prints 1% So let's say Gold is down -3% this month, and that strategy loses -3%. But the Multi-Market Strategy prints 1.6%. Then if they were equally weighted? Then I only lose 1.4%. But maybe they aren't equally weighted. Maybe I have MORE weight to the multi-market strategy, and the Gold position has less weight in the portfolio. Then I might come out, MAKING money. Point being? This is the heart of Non-Correlated Strategies, and is a way to illustrate, what I posted above.
What a brilliant and well thought out reply! Thanks for putting in the time. And so like a Casino, these sort strategies lay outside the realm of us 'one man band' retail traders. The crux appears to be the complexity of multiple programs and strategies which is manageable via a whole team approach. Thank you for an insight into the Quant world. Another strange question I have, after you have been at work, do you come home and work on your computer, perhaps trading? I'm retired now, but in my former positions, some days I would receive over 100 emails of which I needed to read every one and address the majority. Even now, a few years into retirement, emails! Hate them and avoid sending or answering, whereas previously I looked forward to emails.
Thank you for your thoughtful response. As an amateur retail, it is difficult for me to have multiple strategies going all at once as I don't have an army of mathematicians, computer programmers cranking their software day and night to squeeze the 0.1% out of millions of Non-Correlated trades. The best I can do is try to find a strategy with a small positive expectancy to exploit similar to blackjack players counter cards. But that is a tough gig. Anything else we amateur retails can do to take money away from you? Regards,
Imo, best hope for retail is keep away from the domain of intraday, the pros have that covered and retail cannot compete.
You can take money away from him by not going to his "casinos". The only way the casino can loss is if no one shows up.