So maybe you've heard about that term. It refers to the false intuition that gamblers have that if they bet on red for instance, and one or more black comes out, then there is more chance for the next draw to be red (because if so many black come out, certainly it has to be red soon)

it is called "gambler's fallacy" because this is totally wrong of course. Statistically, doesn't matter if 10 or 100 blacks are drawn, there is always the same probability for a red to come out next, and this probability is always 50%.

But what about trading strategies ? Strategies have a higher winrate than 50%. So is "the gambler's fallacy" true as well for >50% strategies ?

well I tested every possible outcome of a strategy, using 10 trades each time:


what we can conclude from this chart is that the winrate is NOT important, what is important is the number of consecutive losses
a strategy with lower consecutive losses will see its winrate improve greatly (if that amount is 1 then the strategy will have a 100% winrate))

so the "Gambler's fallacy" is NOT true for strategies, it depends on more than that
Jeff

ionone wrote: Sat Apr 08, 2023 9:22 pm

So maybe you've heard about that term. It refers to the false intuition that gamblers have that if they bet on red for instance, and one or more black comes out, then there is more chance for the next draw to be red (because if so many black come out, certainly it has to be red soon)

.....reminds me of my visit to Las Vegas many years ago. Went there for a meeting. Late in the evening I decided to walk around in the gambling area of the hotel. I don't gamble, but I find it interesting to watch people doing so. There seemed to be a lot of people with unlimited funds. Anyway, I had doubled my 100 Dollar fund......hahahaha.....after two hours. And before calling it night, I watched one table where black came up 7 times in a row. So I decided to put all my money I had in hand on red. Sure enough black again. And then Red thereafter.......

In trading we need to be aware that a trend can last a lot longer than our account.....i.e. never let thoughts like it must reverse now" cross your mind.

Besides, win/lose ratio is one thing. But we must also have a look at the size of the average losing trades in comparison to the average winning trade because this will determine the required win/lose ratio.
Just my3 Cents. .

ionone wrote: Sat Apr 08, 2023 9:22 pm well I tested every possible outcome of a strategy, using 10 trades each time:

This experiment is erroneous on so many levels that I decided you wanted to crack a joke: the idler's fallacy (one could call it)
Should you ever really want to test something in a statistically relevant way: the Gaussian distribution starts with a minimum of 1000 (as every trader should know who wants to claim that his/her trading system has a such and such winrate).
Happy Easter

josi wrote: Sat Apr 08, 2023 10:18 pm This experiment is erroneous on so many levels that I decided you wanted to crack a joke: the idler's fallacy (one could call it)
Should you ever really want to test something in a statistically relevant way: the Gaussian distribution starts with a minimum of 1000 (as every trader should know who wants to claim that his/her trading system has a such and such winrate).
Happy Easter

the number of data is only dependent on the resolution of the winrate. If I wanted to test 73.4% winrate I would need 1000 datas, but I used steps of 10 % (90%, 80%, 70%) so I only need 10 trades to do so

Jeff

what I need to add is that the number of consecutive losses is interdependent on the winrate of the strategy.
a strategy with higher winrate will inherently have less consecutive losses, but you'll have also less trades in the end :
higher winrate -> GF -> less trades
lower winrate -> GF -> more trade

Jeff