Why do traders accept black swan risk?

September 21st, 2017, 11:19 AM

Why do traders accept black swan risk? Or rather why do black swans exist when they neither benefit traders nor brokers? We can imagine that for short-term traders that black swan risk is not desired because the pay off is too low to develop a winning strategy around and yet the catastrophic risk is still significant over a long enough time. Also, black swan risk is not beneficial to the brokers who will need to try to collect from traders who won't have the money and thus put their own companies at risk of going into default. It seems if something was not desired by everyone then the markets could engineer the risk away.

The only answer I can think of is that the markets aren't really designed for short-term speculators but rather for large traders who need to hold positions for long periods of time and need to use the markets for hedging. And, they are also built for large liquidity providers who can profit from the spread.

In my crypto-currency thread, I speculate a frictionless binary options market could be developed which wouldn't require a spread or market maker but instead would be a parimutuel market. And if you look at the largest trading companies, they all want to be the market maker. They want to provide liquidity. That's very nice of them, right. Why do two "losing" traders need a middle man? The way it would work is that you bet up or down and you do not need to beat the 1% to 5% market maker spread because you get to bet on the true mid point. The only trade off is that your profits are a function of your relative advantage over the other traders. For example, if everyone bets up and the markets go up then everyone gets back their initial bet for a 0% loss and a 0% pay out. If on the other hand, say you bet up and another trader bets down then you get a 100% pay out (assuming they are equal sized bets). This would be a truly zero sum market -- which if actual markets are negative summed -- should be better for short term traders. Even if it wouldn't attract large traders, shouldn't such a market attract volume in the form of small traders because you would be able to bet with zero frictions? (There might be some sort of monthly fee for running the exchange/computers.)

I imagine it might work on a turn basis but more advanced methods might allow for more fluid natural trading. Anyway, one issue might be HFT traders trying to game it by submitting the order at the last second of the open bets period. However, I'm thinking this could be mitigated by using the mid point as the volume weighted average mid point of the bid/ask spread over the time and/or by introducing a slight premium for late bets. I'm assuming you could also do knock-outs, i.e. similar to stop and target, using similar principles.

I am also curious how such principles would work with backtesting. Because, in backtesting you can only test the price but not the ability of the other traders to predict the price. I think as long as you didn't add the tie breaker then backtesting would still be valid because you would be guaranteed to at least not lose money on your winning trades. Though, sure you could end up breaking even on your winning trades and still losing on your losing trades. Also, if you have a large edge then you might be better going to the market maker or liquidity provider because you would be guaranteed a full pay out. However, for short-term day traders who don't really have much of an edge seems like it would work better. The only other catch is that if the average short-term day trader is actually much better then the average market participant, i.e. large trader, then it might prove to be very difficult -- in that case though the parimutuel market would be a leading indicator of the futures market.

If you look at say futures markets, they are skewed to benefit the large trader, either the trader who will hold their position or the large liquidity provider who can sit on both sides of the bid/ask and profit from the spread. I am thinking that a parimutuel market would level the playing field because you do away with the middle man, i.e. liquidity provider. You could either pay a monthly fee to access the market which would pay the datafeed and might provide a seed (reward) value to encourage trading. The trades could either be sealed or open market. In the sealed scenario, you won't know what you can win until after you place your trades whereas in the open market scenario, the pay offs would change in real-time.

1. Sealed scenario. You have to trade in turns. Such as market higher or lower in X time. You have a certain amount of time to place your trades/bets. These are matched against others in the pool. The start value is a weighted moving average of the mid point of the bid/ask spread with possibly some some sort of efficient premium "give up" for late betting. There may or may not be a seed value or minimum win.
2. Continuous scenario. The pay out are updated in real-time. In this way, you can see the aggressor side or dominant side like in traditional order flow markets.

Parimutuel markets might, also, allow for purely algorithmic securities pricing with no need for traders. Shares would be made available for buying or selling based on rationing principle. While not good for those who like to trade, it also gets rid of the middle man and also any irrational pricing. The way it works is you create a programming defined contract which will set the price based on your algorithm. Buyers or sellers can enter at any given turn at the efficient price. More buyers or sellers cannot move the market. Instead, there is a rationing principle based on who gets in first and size. The price is thus guaranteed by the algorithmic pricing but there is no guarantee of liquidity. However, presumably if the programming is rational then there would be buyers and sellers.

September 21st, 2017, 06:47 PM

It is almost like asking: Why do people live in hurricane zones? Sometimes you just have to take the risk...

mattz · September 21st, 2017, 09:11 PM

What do backtesting, market makers, liquidity providers, etc. have anything to do with the fact that traders are willing to accept Black Swan risk?

Black Swan risk is random, unpredictable and no one accepts it because no one can define it, let alone quantify it.
All the disasters we had from the DOT COM, LTCM, and the recent Housing Bubble were a crisis that very few statisticians predict. So you asking retail traders why they accept Swan Risk? They don't. If anything, I feel retail traders have more sense of fear as opposed to the greedy institutions that can run risky models until they are bankrupt.

Matt Z
Optimus Futures

There is a substantial risk of loss in futures trading. Past performance is not indicative of future results.

September 22nd, 2017, 01:02 AM

@Matt

There is a logic to all the ideas I have tried to lay out. First, I have suggested that taking black swan risk is not in the interest in of short term traders. Next, I have suggested that liquidity providers are, also, not needed as one trades on shorter and shorter time frames and, in fact, the liquidity provider or the auction-market system represents a hurdle to overcome. If two short term traders trade against each other, with equal skill, in current system they will both lose. The only way the short term trader can win is to take the other side of a long term trader or trader with less skill. However, in a frictionless system without liquidity providers two short term traders of equal skill could break even against each other.

However, I think the following idea will make it clear as to why traders accept black swan risk: because long term traders need to take the risk because their primary profits are not a function of trading skill. Instead, longer term traders profits are a function of valuation.

In order to understand, let's imagine a business is selling 10 widgets at $10 a piece or has an income of $100. And for simplicity, let's just say their stock valuation is the direct function of this or $100. Now let's say all of a sudden they find a new source of customers in a new market and they are selling 20 widgets at $10 a piece. Their stock will reflect a valuation of $200. However, it never needs to trade at every price in between and importantly the money flow does not need to equal the market valuation. In fact, as soon as this information comes out, any holder will not want to sell for less then $200. So, the stock price will move very rapidly to $200. If the market is efficient very little money will be able to flow into the market from $100 to $200 price. And, that's also why you will see the market make jumps on low volume because every price is an opportunity and is a limited buy. It is just like if you own a house in an appreciating market, you don't need to sale it for the valuation of your house to change.

The markets can be seen as a risk transference mechanism with preference given to liquidity providers and structural advantage to long term traders.

But, the short term speculator doesn't really need to take the risk of market valuation changes because they will probably not be very predictable. Instead the short term speculator is basically trading other traders. They are basically taking the black swan risk to advantage long term traders who use market mechanisms primarily for hedging purposes. It is basically to pay to get access to the primary markets which are most efficient and/or to take the other side of long term traders. However, unlike HFT liquidity providers, retail traders have select disadvantage which is why I suggest a new market structure that is frictionless.

The question regarding the backtesting is say today that if you go on a short enough time frame today, you could develop a winning strategy but one that wouldn't be profitable after commissions and spread. In the new model I conjecture which allows traders to trade without any intermediaries, you can now know that you can trade without any spread and you won't lose money on your winning trades but you cannot ascertain how much you will make because it is a function of your relative advantage of whoever else is in the market trading against you. Also, you cannot be certain that the strategy overall would be profitable even though you know wouldn't lose on your historically winning trades.

Long story short: Markets can revalue without need for traders to make trades which is where the black swan risk comes from. And, most traders take black swan risk because the instruments that traders trade favor accurate tracking over limiting black swan risk. Secondary markets that can limit such risk have higher spreads and/or other hurdles today. The market mechanism I suggest would not have such deficiencies, although it would have unique new deficiencies, and would still probably not attract the larger longer term traders who trade the largest size.

Update:
I would point out something else. Imagine you live in a desirable neighborhood and all of the houses have appreciated well. Here is what is worth understanding: if everyone sold their house then none of the houses would fetch the high valuations. The high valuations are based on a small percentage of houses coming onto market.

Also, I would point out the liquidity provider are not really the adversary. It is the system that prevents two traders for trading without any frictions which is purely engineered to benefit the large liquidity provider and their large long term counter party. The large trader is okay with paying the spread because they have a higher profit factor. However, if someone wants to trade super frequently then they can't overcome the spread but they, also, aren't on a level playing field because the liquidity provider will stack the book. More over, they expose themselves to black swan risk because assuming the profit target will always be filled but the stop loss can be blown. And, assuming black swan risk is a type of instant market movement expressed in efficiency then such movements are more likely to become more frequent. In a perfectly efficient market, you will have the market flat line and then jump instantly to the next level.

Some exchanges have tried to engineer some of the risk away by introducing binary options or other options. However, these introduce tracking error. But also they will have a market maker or a liquidity provider and they will give them an additional spread over the futures market so they will quote the market. So, this makes it even more difficult to profit. But, as I have shown, this is not required as traders can trade directly among themselves once they are willing to accept variable payoff. In such a market, the best trader is more likely to be able to make a profit even if it the size of that profit can't be guaranteed.

I would point out something else too. In the traditional narrative, the liquidity provider is taking risk by providing liquidity. But, wait a second if you take say the ES futures and you have buyers and sellers transacting shouldn't they be able to transact at the midpoint between the spread? Why couldn't you have an order type that says hey match me in the middle. Think about it if 1,000 trades go off then that's 1,000 times the liquidity provider gets to capture the spread. That's 1,000 times both the buyers and sellers were disadvantaged by the spread.

So, what I'm arguing against is basically the futures exchanges is that they somewhat disadvantage the small speculator esp the small speculator who wants to predict the market and trade very actively. Because, the small speculator can't keep enough orders out to capture the spread and is always at the back of the book. Also, the spread prevents the trader from taking the fair zero sum bet. And you notice these liquidity providers want to quote more and more markets because they have the HFT infrastructure to make guaranteed profits.

Hopefully this will wake up the small speculator and/or secondary exchanges to consider offering zero friction trading products.. such as binaries facilitated by parimutuel pool microstructure.

Let's take a look again, at why traders take black swan risk because markets are priced or rather move not based on exchange of dollars but based on rational market valuation. Futures products provide some of the best tracking and lowest fees but still have spreads that disadvantage the small speculator who has a small holding time and can't obtain good book position. Secondary products as exist today have tracking error but also higher fees because they incentivize either the market maker/dealer or liquidity provider to capture fees from trading. The solution is for an exchange to recognize the value in frictionless trading for the active small trader and create this new structure.

Update:

From a post in another thread I made, "
As for the spread, absolutely it is set to disadvantage small speculators who could be transacting at the mid point for friction free trading. It benefits the large liquidity provider (or market maker) who can profit from the amount of volume transacted to offset the jump risk of losses. It benefits the large trader who will hold for longer periods of time and where the spread is a minimal cost of doing business. But, the small day trader who doesn't have HFT advantage to pull orders and doesn't have size advantage to obtain top of book nor the account size to hold for longer periods of time, it is actually a big disadvantage. Obviously it is a game where two great traders can be dimed to death by the middle man under the guise of efficient markets and that is also why markets can be difficult. I suggested in my other thread a way that frictionless trading could take place by using the midpoint and matching whatever orders come in.

I am not the only to realize this. I believe it was Michael Harris from Price Action Lab who did a theoretical study and shown that the spread in say the ES makes it more difficult to profit under theoretical conditions then say the spread in the SPY, which is smaller."

Question then why do small traders, like myself, gravitate the trading futures? The answer is that there doesn't exist a better available option.

Don't get me wrong: I love futures. But, that is just because a superior instrument doesn't exist. NADEX does provide an interesting alternative: however, you have even more frictions. But, here I have already layed out the mechanism for friction free trading.

The perfect trading instrument will be highly granular (small contracts), low frictions or no spreads (direct to direct trade), high liquidity, high leverage, and high tracking.

Stocks have high granularity, low spreads, moderate to high liquidity and low leverage.

Futures have low granularity (large contracts), large spreads, high liquidity, high leverage, and high tracking.

Binary options have high granularity, higher frictions, and lower liquidity.

Parimutuel pool trading which can be binaries or knockouts or other futures like instruments can be designed to have high granularity (small contracts), near zero frictions, high leverage, moderate tracking, and modest liquidity (but should grow once traders understand the benefits).

Michael Harris' simulation https://www.priceactionlab.com/Blog/2015/04/the-perils-of-day-and-position-trading-in-zero-sum-markets-part-one-future-day-trading/

September 25th, 2017, 08:56 PM

Not much interest? Ideas welcome on how it would be possible to create frictionless markets without black swan risk. It looks like Bitmex? has perpetual swaps on Bitcoin. I'm curious if this could work with futures.

There's something peculiar too in futures markets: you can't cross the spread. You can't match in the middle (to my knowledge) or else I don't know of such order type.

One idea with the parimutuel pools betting I've had is you take turns and if done properly this might slow the market down and limit the HFT advantage. Basically you go long or short at the mid point. However, the next tick is determined based on whether more people go long or short. So, if you can predict the market ticks up you will be in profit. This is similar to how the markets trade today except you don't have to pay a premium.

I'd like to see a market structure that:

1. Allows as close to 50/50 betting as possible.
2. Neither advantages larger traders or disadvantage larger traders
3. Minimizes HFT advantage.
4. Does away with advantage for "locals", market makers, and liquidity providers.
5. Prevents black swan risk.
6. Tracks very closely to the underlying market.

I'm thinking all the orders get thrown into a black box. The order type is basically FOK. You won't know if it fills until the next turn.

I'm surprised the CME is not interested in ideas like these. If they would provide the right market microstructure, us retailers could trade a lot more frequently. Maybe its not the primary market but a new product for small traders.

I am thinking that in the past the futures were better trading products because the trade fees were really high. So, the large tick size sorta offset that. But, now you can trade stocks commission free with minimum of spread. Why should someone trade ES when they can trade SPY/ETFs for free? i.e. at places like Robinhood.

One possibility would be a "call-able" future. The way this might work is that say I have $200 for my futures trade. I back it with $200. Someone takes the other side with say $500. And let's say it gets "called". In this case, it will be automatically closed. What about the counter-party though? There trade would get some sort of benefit to exchange or match or reopen without a cost. This is just an a rough idea but makes more sense then taking black swan risk.

September 26th, 2017, 10:42 AM

Interesting concept. I moved from holding overnight to day trading to minimize black swan risk. Doesn't eliminate it but reduces it. Some of what you suggest can be accomplished by taking options positions instead of spot or futures. I have considered that but I am not happy with the fills in the futures option markets.

Sent using the NexusFi mobile app

Why do traders accept black swan risk?

Discussion in Traders Hideout

Why do traders accept black swan risk?