The 80% rule

[speedo]

I've never seen 80% in trading, not even in backtests (which probably counts as a good thing, the system was correct / honest). But that's a different thing, 80% real life bets DO exist.

80% is easy to achieve....if you're willing to eat like a bird and shit like an elephant.

 

[Handle123]

I've never seen 80% in trading, not even in backtests (which probably counts as a good thing, the system was correct / honest). But that's a different thing, 80% real life bets DO exist.

Most likely you never will, what would anyone who has them to gain by telling anyone?
I have had up to 95% losing commodity trades in a year and did well cause all the hedging, so you can do miserable in one area and clean up in other.

 

[Aquarians]

Most likely you never will, what would anyone who has them to gain by telling anyone?
I have had up to 95% losing commodity trades in a year and did well cause all the hedging, so you can do miserable in one area and clean up in other.

Yeah, but unless I was drunk or otherwise don't remember well, you also claimed in the 98%+ positive on the WINNING side. Which leads us to your original question.

 

[Metamega]

I think you’d have to be naive to think in a competitive marketplace that any profitable edge with 80 percent win exists.

 

[Palindrome]

Virtu Financial, one of the largest HF Trading shops:

Per trade 50% win rate, 25% lose rate, 25% scratch rate

Yearly 99% of their days are green

 

[wildchild]

As traders I'm willing to bet you balk at strategies with low probability of success. Yeah I know, what matters is expected value, win = probability * amount so if amount is big enough you're still profitable as long as your repeated investments are low enough so you don't go bankrupt before realizing that profit.

Still, I by far prefer 80% probability of success over 1% and am sure that also do almost all of you.

And when probability = negligible and amount = ridiculous(ly low), no sane trader would take it. Yet, applying to jobs = 1 in 10,000 applicants probability of success, where success = a measly salary. All while investing a fuckton of money (time) on each "trade" (job application). Automated code tests which take hours of your time while costing the employer practically nothing. Worse, "homework" tests which cost you days.

From now on I've got a rule, I'm no longer placing trades with lower than 80% probability of success. And by "trades" I mean a broader interpretation.

More details later.

That is not going to work. Putting together a 80% of success strategy is not that hard. The problem is, the 20% of the time it fails, you are going to get flattened.

 

[Palindrome]

That is not going to work. Putting together a 80% of success strategy is not that hard. The problem is, the 20% of the time it fails, you are going to get flattened.

It's a myth all traders want to believe is true. Win all the time, lose infrequently. It's pain avoidance.

We are programmed to want this to be reality.

We are programmed to rather have steady gains, then "lumpy" gains...
EVEN IF the lumpy gains are larger then the steady gains...we still would rather the steady...again because we are programmed this way.

 

[maxinger]

Archers aim for 100% accuracy.
Surgeons aim for 100% success.

Traders are different.
100% or even 80% success rate is a very bad lousy system.

I aim for 50 to 60% success rate.



This is probably the result of high success rate system :

+5 ticks
+2 ticks
+ 3 ticks
+ 4 ticks
+ 1 tick
- 100 ticks
+ 4 ticks
+ 2 ticks
- 80 ticks
+ 2 ticks
ie eat like mouse, shit like elephant.


This is my typical result :
+ 50 ticks
- 15 ticks
- 12 ticks
- 1 tick
+ 5 ticks
+ 20 ticks
- 15 ticks
+ 2 ticks
+ 80 ticks
- 12 ticks

 

[smallfil]

As traders I'm willing to bet you balk at strategies with low probability of success. Yeah I know, what matters is expected value, win = probability * amount so if amount is big enough you're still profitable as long as your repeated investments are low enough so you don't go bankrupt before realizing that profit.

Still, I by far prefer 80% probability of success over 1% and am sure that also do almost all of you.

And when probability = negligible and amount = ridiculous(ly low), no sane trader would take it. Yet, applying to jobs = 1 in 10,000 applicants probability of success, where success = a measly salary. All while investing a fuckton of money (time) on each "trade" (job application). Automated code tests which take hours of your time while costing the employer practically nothing. Worse, "homework" tests which cost you days.

From now on I've got a rule, I'm no longer placing trades with lower than 80% probability of success. And by "trades" I mean a broader interpretation.

More details later.

What matters is your expectation or edge. If you have negative expectation even a 90% win rate will not matter. The 10% of the time you are wrong, that would be enough to wipe you out totally! In contrast, a 30%, 40% or 50% with a huge edge will result in a positive expectation enabling you to win enough to wipe out all those small losses. You need large gains and small losses when you have a lower win percentage.

 

[tsznecki]

And when probability = negligible and amount = ridiculous(ly low), no sane trader would take it. Yet, applying to jobs = 1 in 10,000 applicants probability of success, where success = a measly salary. All while investing a fuckton of money (time) on each "trade" (job application). Automated code tests which take hours of your time while costing the employer practically nothing. Worse, "homework" tests which cost you days.

Oh I don't know because a salary pays for food and rent? Did the automated code tests magically write themselves?

Either easy shit like "web" or, perhaps counter-intuitively, very hard stuff like "quant finance". With the easy stuff it's intuitive to figure out there are a lot of people able to qualify to the level required to do the job. With harcore stuff, the problem's in the number of jobs available which is very low actually. Much lower than the number of very capable individuals who have no alternative in this shitty crappy sucky world we live on, but to compete for those few positions which offer something more intellectually stimulating and financially rewarding than shoveling shit, which is where the "optimum" is: job position exceeds "qualified" (and willing to take the crap) individuals.

Oh web is easy? Let's just bring that up to the engineers at Google, Facebook, any Big N tech firm.

So you have 100 Einsteins, Wisensteins, Abakumovs etc and ONE fucking job which allows them to unfold their talent. What, you never heard of Wisenstein and Abakumov? Well it's THEIR fault for not getting the job which rescues them from immediate extermination or at best, a life in the Gulag. Where after 14+ hours of meaningless shoveling crap work excluding chores, they should have proven themselves if they were any good anyways.

Life is hard. You think you are entitled to something? Fucking loser.