Today, I did something I have not done before: side by side comparison of real time live display of Schwab stock price and TOS paper trade "live" display of PA. Guess what? The sim PA lagged a few seconds behind real time live data. What that means is some YouTube super traders could show "live" trades and be able to game the system and produce outsize profits?
Good Morning ironchef, You are definitely correct. A lot of YouTubers selling trading courses and trading live is lying and cheating to get more trading courses sales. There is a big difference between selling trading services products and trading to make money. It hard to believe anything sellers are saying.
If you look carefully at my OP chart, this is exactly what is happening: The change was gradual, win rate went up, R:R down, profit % down and the final outcome is ==> gradually losing.
And yesterday I added this to my trading scheme: The result, adding the above feature to one trend changed the day from a loss to a winning day. However, one data point doesn't prove anything yet.
I checked again today, comparing sim and real time prices. Today, the lags in sim price is almost non existence. Very strange.
Good Morning ironchef, Lol, are you thinking of doing something "strange for some change"? You know the old saying "If you can not beat them, join them" hint hint hint
Aloha @Laissez Faire ! I've started doing a lot more stats work, inspired by some of your posts. What types of days do you classify each as (Trend, Range, etc?) and by metrics do you use to classify a day's structure?
Glad to be of inspiration. Regarding day types, I'm not sure I want to share any specifics there, but I'm sure you'll find some repeatable structures that could be worthwhile to classify in your catalogue of day types. At the most coarse level, you'll have the Up Day and the Down Day - going from low to high or high to low. Good luck.
Thank you for the reply. Maybe I can milk this more vague answer outta ya: How many types of days do you think there are to classify? (Trend + Range = 2. Trend + Range + V = 3... and so on) I see there being 6 types: Trend, Range, V, Lightning bolt (up/down/up or down/up/down), M, and W. Do you think that's too many to use to classify?
Depends what you're trying to explain What information does each day provides ? Information is the difference from the most basic mean / median. Let's say ... the median daily (close - open) is +5.9% of the daily range. +5.9% is our most basic mean / median. It's our benchmark. Now ... We know that yesterday's close > yesterday's open. Does it provides any information ? We try it ... The median daily (close - open), knowing yesterday's close > yesterday's open is +4.6% of the daily range. Which means that knowing the previous (close-open) provides some information ... An up day "decrease" the next day's (close-open) median by 1.3% of the daily range. The median daily (close - open), knowing yesterday's close < yesterday's open is +7.6% of the daily range. A down day "increases" the next day's (close-open) median by 1.7% of the daily range. As long as your types of day provide information (Decrease uncertainty) then it's fine. Now that’s the most basic stuff. If you want to “predict” what type of day it’s going to be then you need to find “priors” or filters that will “decrease the uncertainty” from the initial maximum uncertainty (1/n). I think the most basic is range or trend day. Already knowing that is going to be an edge. Then you have trend up, trend down, range. Gives you even further edge. If you want to know W or M day then you’ll definitely need intraday data and find reliable priors. But these are just statistics and we know the unlikely will happen. Then we need to build a strategy around the statistics that earn money in the long run. My 2 cents.