If it takes time to learn how to trade profitably, those initial losses would probably counteract later gains in such a way that even an experienced, successful trader would have been better off sticking with SPY from the start.
A way around this would be for rookies to keep 90% of their portfolio in an index fund and play around with 10% silly money.
If after a decade or so the proto-trader reckons he's ready for the big time, he might increase the amount of funds he plays with. If he finds the results disappointing (more likely), the losses will be limited and he will have painlessly learned his lesson.
I think most people trade or pick stocks just to prove to themselves or to finance Twitter bros how clever they are. I credit all my success with money to admitting that I'm an idiot.
True. I think the advice to stick to SPY prevents many an inexperienced trader from burning their fingers needlessly. But with time, and by allocating a small portion of your portfolio to other strategies, you might be able to find an edge which is responsible for outsized returns. After all, long term returns cannot be cashed out now!
Figuring out how much to allocate to your various strategies is a really important skill - I might have a few things to say on that soon. :)
Feb 13, 2022·edited Feb 13, 2022Liked by Market Sentiment
I've been bench trading for several months, with your sentiment tracking tool as one of my inputs, and I'm averaging a little under 1% per day over S&P 500 on ≤24-hour trades and about 1.5% on 24-48 hour trades. Compound that over 252 or 126 trades per year and the results are tremendous.
It's not straightforward, mind you: I track (currently - always expanding) four different buy/sell timing strategies and eight selection strategies, and have created an analytics platform to test them and optimize. Preliminary testing of optimizations in progress suggests I can increase those 1- or 2-day returns by a further 0.5% - 1%.
There may be opportunities to improve your sentiment tracker, which would further increase its predictive power. My suggestions, if not already implemented:
• Weight your sources according to the reputation of the poster, ex. Reddit Karma, Twitter followers, etc.
• Weight your sources according to the positivity of the response to the post, ex. Reddit upvotes, Twitter likes, Twitter retweets, etc. I suggest some combination of the total positivity and the positivity ratio (ex. likes per follower).
• Create a weighting factor that correlates sentiment with stock price movement. This will reduce the influence of the meme stocks and the stocks users "can't let go of". My inferences are that: 1. users who invested in meme stocks can't accept their sunk costs and move on, so they keep discussing - with undue optimism - their poor investments; 2. stocks with high beta values were frequent "winners" during the extreme bull market of the past two years, and users may have difficulty letting go of these stocks in the (probable) upcoming bear market.
Thank you for your sentiment tracker and your newsletter! Excellent resources and I look forward to where they can go in the future.
Wow! It's great to hear positive feedback on the tracker. I had stopped working on it actively to focus on the newsletter, but your suggestions are great. Weighting tweets and posts by influence was on my to-do list. Will get around to it in the next round of developments.
(For those who are unfamiliar with the tracker, here's the link: https://marketsentiment.live. Play around with it to track real-time chatter and sentiment around different stocks!)
It's interesting to see how you have applied it, gives a clear indication of what works and what's useful. Thanks!
You created a valuable tool, so the least I can do is share my appreciation! And helping with the algorithm benefits us both :) It would be interesting to better understand your tracker if you're comfortable sharing your data sources, algorithm, etc., but it's understandable if you prefer to keep it proprietary.
My interest has always been more with data than stocks, so this started as a quick test to see whether a couple of free stock sentiment trackers had any predictive power. It seemed plausible that:
• Randomly selected stocks should generate returns, on average, equal to the market (accounting for weighting by market cap., representation in various indices, etc.)
• Timely and insightful sentiment tracking should reflect either a true insight into performance or baseless hype, either of which could be a precursor to a price increase
• Delayed or reactive sentiment tracking could lead to negative performance if it registers sentiment after the price has already moved and/or picks up on chatter that is simply a reaction to price movement
Several tools and strategies led to the latter scenario. Over time, I've been refining my inputs and methods and performance has been improving. Obviously, a better sentiment tracker will lead to better performance and I'm happy to do whatever I can to assist.
A note about poker is that when you look at the top 5, say, then we see a much larger share of repeat winners. The question of whether we should only look at the top 1 Vs top 10 isn't an easy one either. If you did chess and randomised a few of the movements or rules I wonder if we wouldn't see similar charts atop.
True, the top 5 is a good metric and even tracking the amount won/lost would be interesting. The comparison was probably a little unfair because chess is a winner-take-all game while Poker is more about asset allocation.
As an aside, the connection between the worlds of gambling and trading is quite interesting - The first thing that comes to mind is the speculative aspect, but there's a lot more to it: Probabilistic thinking, asset allocation, a combination of chance and skill... Will definitely be looking into it more!
Trading is a discipline in its own right that requires very specific skills. The problem is that very few people have these skills. Many will still want to trade and lose a lot of money because they will be up against experienced traders who have more experience and skill.
If you don't have these skills, your best option is to use your greatest natural advantage: the time you are willing to spend in the market. Playing the long game is your best option.
If it takes time to learn how to trade profitably, those initial losses would probably counteract later gains in such a way that even an experienced, successful trader would have been better off sticking with SPY from the start.
A way around this would be for rookies to keep 90% of their portfolio in an index fund and play around with 10% silly money.
If after a decade or so the proto-trader reckons he's ready for the big time, he might increase the amount of funds he plays with. If he finds the results disappointing (more likely), the losses will be limited and he will have painlessly learned his lesson.
I think most people trade or pick stocks just to prove to themselves or to finance Twitter bros how clever they are. I credit all my success with money to admitting that I'm an idiot.
True. I think the advice to stick to SPY prevents many an inexperienced trader from burning their fingers needlessly. But with time, and by allocating a small portion of your portfolio to other strategies, you might be able to find an edge which is responsible for outsized returns. After all, long term returns cannot be cashed out now!
Figuring out how much to allocate to your various strategies is a really important skill - I might have a few things to say on that soon. :)
I've been bench trading for several months, with your sentiment tracking tool as one of my inputs, and I'm averaging a little under 1% per day over S&P 500 on ≤24-hour trades and about 1.5% on 24-48 hour trades. Compound that over 252 or 126 trades per year and the results are tremendous.
It's not straightforward, mind you: I track (currently - always expanding) four different buy/sell timing strategies and eight selection strategies, and have created an analytics platform to test them and optimize. Preliminary testing of optimizations in progress suggests I can increase those 1- or 2-day returns by a further 0.5% - 1%.
There may be opportunities to improve your sentiment tracker, which would further increase its predictive power. My suggestions, if not already implemented:
• Weight your sources according to the reputation of the poster, ex. Reddit Karma, Twitter followers, etc.
• Weight your sources according to the positivity of the response to the post, ex. Reddit upvotes, Twitter likes, Twitter retweets, etc. I suggest some combination of the total positivity and the positivity ratio (ex. likes per follower).
• Create a weighting factor that correlates sentiment with stock price movement. This will reduce the influence of the meme stocks and the stocks users "can't let go of". My inferences are that: 1. users who invested in meme stocks can't accept their sunk costs and move on, so they keep discussing - with undue optimism - their poor investments; 2. stocks with high beta values were frequent "winners" during the extreme bull market of the past two years, and users may have difficulty letting go of these stocks in the (probable) upcoming bear market.
Thank you for your sentiment tracker and your newsletter! Excellent resources and I look forward to where they can go in the future.
Wow! It's great to hear positive feedback on the tracker. I had stopped working on it actively to focus on the newsletter, but your suggestions are great. Weighting tweets and posts by influence was on my to-do list. Will get around to it in the next round of developments.
(For those who are unfamiliar with the tracker, here's the link: https://marketsentiment.live. Play around with it to track real-time chatter and sentiment around different stocks!)
It's interesting to see how you have applied it, gives a clear indication of what works and what's useful. Thanks!
You created a valuable tool, so the least I can do is share my appreciation! And helping with the algorithm benefits us both :) It would be interesting to better understand your tracker if you're comfortable sharing your data sources, algorithm, etc., but it's understandable if you prefer to keep it proprietary.
My interest has always been more with data than stocks, so this started as a quick test to see whether a couple of free stock sentiment trackers had any predictive power. It seemed plausible that:
• Randomly selected stocks should generate returns, on average, equal to the market (accounting for weighting by market cap., representation in various indices, etc.)
• Timely and insightful sentiment tracking should reflect either a true insight into performance or baseless hype, either of which could be a precursor to a price increase
• Delayed or reactive sentiment tracking could lead to negative performance if it registers sentiment after the price has already moved and/or picks up on chatter that is simply a reaction to price movement
Several tools and strategies led to the latter scenario. Over time, I've been refining my inputs and methods and performance has been improving. Obviously, a better sentiment tracker will lead to better performance and I'm happy to do whatever I can to assist.
A note about poker is that when you look at the top 5, say, then we see a much larger share of repeat winners. The question of whether we should only look at the top 1 Vs top 10 isn't an easy one either. If you did chess and randomised a few of the movements or rules I wonder if we wouldn't see similar charts atop.
True, the top 5 is a good metric and even tracking the amount won/lost would be interesting. The comparison was probably a little unfair because chess is a winner-take-all game while Poker is more about asset allocation.
As an aside, the connection between the worlds of gambling and trading is quite interesting - The first thing that comes to mind is the speculative aspect, but there's a lot more to it: Probabilistic thinking, asset allocation, a combination of chance and skill... Will definitely be looking into it more!
Trading is a discipline in its own right that requires very specific skills. The problem is that very few people have these skills. Many will still want to trade and lose a lot of money because they will be up against experienced traders who have more experience and skill.
If you don't have these skills, your best option is to use your greatest natural advantage: the time you are willing to spend in the market. Playing the long game is your best option.