I've manually tracked my discretionary setup by collecting over 100 screenshots across three instruments. Reviewing them, it seems to win about 60% of the time, with winners often giving 2R or more. Since it's not a coded or rule-based system but more context-driven, I'm wondering — is this a valid way to estimate expected value? Or does EV only become meaningful when it’s tracked in a structured database or spreadsheet? I’m trying to understand where discretion meets quantification.
Very interesting information, probably the most useful I’ve read in a long time. More articles like this, please. I have a question about the sample size. In the text, you mention: "For meaningful probability estimation, a minimum of 30 instances of a specific setup is required, with 50+ instances preferred." Isn’t that a bit low? Why do some traders claim that you need 300 or more for a proper backtest?Yes, a backtest is something different from probability estimation, but still.
You're absolutely right to question this—30-50 is just the absolute minimum threshold, not optimal. Serious backtesting demands 300+ samples for statistical validity. My article established the floor, not the target. More data always yields better confidence in your probability estimates.
Im not the sharpest tool in the shed so sorry if this is obvious. When doing the EV formula, where is the number you subtract the result of the equation coming from? In your example its the 4.4 you subtract from the 9.
I've started trading 2 years ago. I am self learner. I've gone from learning form YouTube trading gurus to yourself and Trader Dante studying materials and I cannot stress enough how much of a difference it makes. Breaking down key elements that are so often overlooked. I really enjoyed this post and I found it very informative. Will you ever write a post/tutorial/step-by-step on your personal approach of collecting all the data?
An extremely informative post, thanks Ryan. I’m beginning to see now, how one needs to reconstruct their journaling to truly define and track their edge!
In the article, you write the following: "For meaningful probability estimation, a minimum of 30 instances of a specific setup is required." I would like to ask what you mean by 'specific setup'? Could you provide an example? It's a great article, really. Thanks for sharing
I've manually tracked my discretionary setup by collecting over 100 screenshots across three instruments. Reviewing them, it seems to win about 60% of the time, with winners often giving 2R or more. Since it's not a coded or rule-based system but more context-driven, I'm wondering — is this a valid way to estimate expected value? Or does EV only become meaningful when it’s tracked in a structured database or spreadsheet? I’m trying to understand where discretion meets quantification.
Thanks again for such great posts.
Brilliant content.
Thanks XO 🤝
Great article. Think there should be a plus sign between the parentheses in the simplified EV formula.
Very interesting information, probably the most useful I’ve read in a long time. More articles like this, please. I have a question about the sample size. In the text, you mention: "For meaningful probability estimation, a minimum of 30 instances of a specific setup is required, with 50+ instances preferred." Isn’t that a bit low? Why do some traders claim that you need 300 or more for a proper backtest?Yes, a backtest is something different from probability estimation, but still.
You're absolutely right to question this—30-50 is just the absolute minimum threshold, not optimal. Serious backtesting demands 300+ samples for statistical validity. My article established the floor, not the target. More data always yields better confidence in your probability estimates.
Im not the sharpest tool in the shed so sorry if this is obvious. When doing the EV formula, where is the number you subtract the result of the equation coming from? In your example its the 4.4 you subtract from the 9.
Thank you for your post, very informative.
Great question!
4.4 is simply the expected loss.
In the example, you win 45% of the time and make +20 points → 0.45 × 20 = +9
You lose 55% of the time and give back 8 points → 0.55 × 8 = 4.4
Because a loss is negative, that second piece comes in with a minus sign:
EV = +9 – 4.4 = +4.6 points.
So the “4.4” you see isn’t subtracted afterwards—it is the loss component (0.55 × –8 = –4.4).
I figured writing “9 – 4.4” is just cleaner than “9 + (–4.4).”
Hope that makes sense!
Thank your for the clarity! you rock
This is great, Ryan. It helps a lot. Will you discuss other common metrics e.g., MFE, MAE, etc. in the future?
Yes, I’ll be covering a lot of ground! Including metrics
I've started trading 2 years ago. I am self learner. I've gone from learning form YouTube trading gurus to yourself and Trader Dante studying materials and I cannot stress enough how much of a difference it makes. Breaking down key elements that are so often overlooked. I really enjoyed this post and I found it very informative. Will you ever write a post/tutorial/step-by-step on your personal approach of collecting all the data?
Siema, dodałem Cię na discordzie mordko!
siema, fajnie widzieć kogoś swojego tutaj. złapiemy się jakoś na discord? dodaj mnie do znajomych: wrotkarz
Thanks for giving us the map for a trading career
An extremely informative post, thanks Ryan. I’m beginning to see now, how one needs to reconstruct their journaling to truly define and track their edge!
Thanks, James! 🤝
In the article, you write the following: "For meaningful probability estimation, a minimum of 30 instances of a specific setup is required." I would like to ask what you mean by 'specific setup'? Could you provide an example? It's a great article, really. Thanks for sharing