Is it important to define your edge mathematically? If so, how do you go about finding one?

[alistera]

I already run a system that does stat arb, it runs on Crypto because it's simpler than standard asset classes, here are the numbers it produced the past weeks.

This can just about be done via manual trading.

This one requires full auto trading which the system does.

No one can do stat arb today without the above, you have to take each and every trade without exception as the system limits losses to fractional while the wins are spectacular usually on a 10:1 basis, missing the 1 isn't good, which is why people keep coming to me asking to implement it for them and we just switched it to a percent profit model, it's used to supply signals to providers and crypto funds mostly (sometimes traders can get discounted access), apparently there are 14 funds looking at it for implementation so might have some jobs for grads in the coming weeks and months.

Actually there's no such thing as an edge but I'll indulge everyone's view, the only quantitative edge you can have today is AI and data and algos (this is what the arb above does), it processes 10s millions of data points per day to come to it's conclusions using institutional fintech, sure people can produce something with a lot of work but those subsystems will not produce meaningful profits and never will going forward, those days are past.

The qualitative edge is using time, the more time you have the better the choices and trades which allow you to invest when everyone is liquidating, today that has now moved beyond VHNW and is on its way to UHNW, all HNW of the world and I know quite a few of them, they are daily losing their net worth and not replacing it, mid-level management are in even worse shape today.

What do we do for trading and investing, we actually do both, we use massive amounts of data and AI based algos built in to the fund management platform, plus we wait longer than anyone for optimal trade and investment setup, these two combined mean your profit basis isn't the usual 1-2% per month but can extend 2x 5x 10x that number at least for investment grade, if you only use one of the edges you will fight a mostly losing battle to keep 1-2% per month, unfortunately very few, like no one, can do both edges without algo support.

You have "two" edges in the markets, actually you only ever did have two but the world was inefficient decades ago, today they are not inefficient so the spread widens, which is where our in-house 10,000pages of unique content on the financial markets comes in to play, almost all financial knowledge is built on the past decades (very narrow) spread of data and time, our knowledge base and platform is built on the current spread which has widened to spectacular levels.

If you can define your edge using data and AI and algos plus also define your edge using time, the world is your oyster, if you can just define technical edge it will flash fail, if you just define fundamental edge it will progressively be better but on an exponential curve which is slow, measured in years, if you can define neither which is most people today, and interestingly the hole HNW are putting themselves in to at an ever faster pace, then you will just burn capital 10x to 100x faster than you can make it.

 

[TrAndy2022]

two questions

1. Is it important to define your edge mathematically?
   
   
2. How do you go about trying to find an edge?

1. No it is not required but it can be sufficient. Most important here that you describe your edge in clear rules which only leave minimal room for further interpretation.

2. Usually by observing the markets. Here you can generate a lot of ideas which you can test them, usually via coding and backtesting, but you can also do a heuristic graphical approach, especially for the beginning, to not overcomplicate it. Second, you can read scientific studies, a good start here, is Quantpedia(.)com with full of good papers on already discovered edges.

Last, it takes a lot of time to be proficient in what you do and you need to spend a lot of efforts to be consistent profitable. So plan you need at least a few years until you get there. And start small because loosing is part of the game too, especially in the beginning of any trading paths.

 

[SunTrader]

OK. There is no edge .

I gotta try that next time. Let someone think they out-smarted you.

 

[PrestonRiver]

I think it is not necessary, but it can help you if you bring mathematical solutions as well.
In the end, your profit or loss will tell you how good your edge is.
and I think you can find it using your ideas; there is no specialized way to find and develop it; just do not underestimate your ideas because they might be the origin of a good edge!

 

[longandshort]

two questions

1. Is it important to define your edge mathematically?
   
   
2. How do you go about trying to find an edge?

1. Very. Need to know your hit rate in order to know how confident you should be.
2. Study. Spend time researching market anomalies until you find some you can trade.

There are roughly 3 types of edges:
- analytical edges, such as having a better way to value a stock
- informational edges, such as knowing a company is about to miss earnings
- behavioral edges, such as avoiding heuristics and logical fallacies (such as sunken cost fallacy) or by being contrarian

Many “traders” are fooled by randomness and don’t have a process that’s truly accretive to pnl because they 1) don’t understand the market and 2) don’t have an edge.

Spend time to study the market not by looking at charts but by reading & analyzing things like how auction markets work, LoB in the age of hf mm, investment strategies of mutual funds, etc. Do this first and then start looking for opportunities.

 

[speedo]

You answered your own question, you discover and define your edge by deriving an edge from the data of your testing and incorporate that edge into a trade plan. If you can't define an edge mathematically, how do you know if an edge exists?

 

[Laissez Faire]

I think many would-be-traders don't understand expectancy. Simply put, if you risk the same as you gain to win, you need a high win rate. If you risk twice what you gain to win, you need a very high win rate. If you risk less than you gain to win, your win rate can go below 50 % and even much below if your taking home high multiples of initial risk.

In practice, most would-be-traders don't have a high win rate and usually cut their profits short afraid it's going to go back to even while taking full losses. The end result is a negative expectancy.

So, understanding this basic math that no trader can escape is a good starting point.

As for finding an edge, I think that depends on what you want to do. Common for all approaches would be a ton of studying. I wouldn't recommend it as you'll most likely end up going nowhere meaningful on your own as a retail trader. That's not meant to be cynical, just being realistic.

Even with a profitable strategy in your hands you still need to master yourself. In many ways, the trading game is rigged against human nature.

 

[TrAndy2022]

There is only one true mathematical edge which will last forever or is the last edge that would disappear. That is compounding on Euler figure. You can use a trailing stop to scale in when you are in profits. That way you can make for a standard 3:1 Reward-Risk-Trade a 2.7^(3)=~20:1 RR trade. In that way you can make much more money than loosing it, without having a big edge. You only need to scale in when you are in profits and a (fixed points or same distance as percentage) trailing stop. Best way here is to make an excel sheet as I have done it, to calculate how much you can add to your winning position without increasing the initial risk. So you basically have the same initial risk through the whole trade but you can can gain multiple times more because you increased your overall position on small steps when you made more profit within that trade. Just for imagination that a good 7:1 RR trade would translate with scaling in on small steps with the law of compounding on Euler Figure which is 2.71828, here ~2.72^(7)=1100 RR times your initial risk as profit what you could make here. The law of compounding is also described elsewhere as the 8th world wunder according to Albert Einstein and other famous people. This is a true mathematical edge you can always use in your favor. Just get familiar with it. This I can recommend to every (new) trader as it is very powerful concept!

 

[Global OptionsTrades]

two questions

1. Is it important to define your edge mathematically?
   
   
2. How do you go about trying to find an edge?

Probably better to look at those the other way round.

2. Study, learn, model, backtest, refine / iterate. Repeat forever.

1. If you can determine that you have a positive expected risk-adjusted return after costs (including your time) you probably have an edge. If you cannot, go to q.2

 

[Sekiyo]

OK. There is no edge .

Jesse Livermore, Richard Wyckoff, Gerard Loeb, Nicolas Darvas, William O'Neil, Qullamaggie, Mark Minervini, Richard Dennis, David Ryan, Stan Weinstein, Dan Zanger, Crypto bro billionaires ... They are all mistaken.

You can't build a fortune buying leading stocks, breaking out, in a bull market.
The only way to make money is to grow 1% per day. 1ES point à la SimpleMeLike