An Argument Against Posting Yearly Results
Posted by Tyler Forrester
Posted by
Tyler Forrester
posted in
Run It Once Training
An Argument Against Posting Yearly Results
(Cross-Posted) Full disclosure of winrates has reverse implied odds for the coaches and likely is a fairly weak indicator of coaching strength, because over a yearly sample, we can't statistically distinguish between 1bb and 4bb winrates and bumhunting has an out-sized effective on winrate.
There are three results from showing my yearly results:
1) I have a spectacular winrate. Now everyone is focused on how to reverse engineer my game, like people dissect Linus or RedBaron. This is clearly a net negative for me as a player. Additionally, if the winrate is due to game selection, I will get more competition in my games. This can far offset any income boost I get from coaching.
2) I have a low winrate. Now even after many years of success, players don't want to watch my videos or get private coaching, because I'm a dinosaur. The thing is that every single player will have a bad year at some point in their careers, because a 1bb winrate and 6bb winrate can't be statistically distinguished from each other over most yearly samples. Additionally, game selection has an absurd impact on winrate (but not hourly), so Runitonce would actually be selecting for coaches who mega-bumhunters.
3) I have Goldilocks winrate, it's not too high and its not too low, then people are content with my ability to play and coach, but aren't going to copy my entire approach to poker.
The only scenario where I win is if I hit the Goldilocks zone. Otherwise, its -EV to post.
Loading 7 Comments...
How else we can measure coaches? By quality of content? Thats not simple either.
To paraphrase John Wanamaker: "When I get coaching half my money is wasted, if only I knew what half."
Any teacher is going to provide different EVs to different people which means there is no hard metric to evaluate a coach's value. Learning has 3-parts, the teacher, the way the information is communicated, and the student. A teacher has to be knowledgeable about the subject, but the student's EV depends on how well the student can understand the teacher; how quickly the student can learn; and what the student already knows.
For example, I like to learn from players, who win but have obvious leaks in their flop and preflop game, because this means that they are doing something better than me on the turn and river.
Another +EV coach for me is a player who wins but plays very differently, because he gives me new insights into possibly profitable lines. The diversity argument is strongly supported mathematically, so if you want the best strategy, you need to have a way to choose the best the parts out of many strategies. One common approach is to take the middle-of-the-road strategy by having each strategy "vote" on play and then choose the play that is recommended by the most strategies.
I can guarantee that novice or intermediate player will want something different. Maybe they will want a thought process to guide new decisions or a hand-chart to help with preflop mistakes, but certainly not learning two or three strategies to find commonalities or learning from players who demonstrably play poorly in some situations, but still win.
TLDR: Coaching value changes based on both the coach and the student.
I assume you are talking about mid-high stake reg pool.
Well, those players knows the game pretty well along with the built-in variance, so in theory, results shouldn't be the only metric.
If one competent player take coaching from a crusher (who have ok results) and from a bogus one (with astonishing winrate) he can easily detect the discrepancies and instaquit the latter.
The problem is at the micro-low stakes where a player cannot distinguish properly a legit from a quasi-scam coach
These are valid points, another thing to consider is that even someone with a spectacular win-rate might not be able to communicate concept and strategic advice well. Where a break-even rakeback grinder might give valuable content and might give more value to student's. It's all about quality of content and what the audience does with the quality. I've watched coaches that might not have had spectacular results but that simply opened my eyes about concepts and intricacies about the game that I did not realize before which can be very valuable.
I agree, I don't think it's necessary to post the results. I imagine it also adds a tension that undermines your enjoyment of the game and that can affect the entertainment value of your videos. What does matter to me is that the coach provides a unique perspective, either by thoroughly teaching to their strengths, or exploring their weak spots.
A video I'd like to see from a coach is one that looks for solver GTO strategy deviation from the current player population's meta.
Another thing to consider is also that it is ridiculously easy to doctor your results by deleting losing stretches etc etc. With no real way of verifying data how much worth are these stats ? I believe most coaches here are honest but I'm just saying that its really easy to manipulate your results if you want to
In your case Tyler, I couldn't care less of your results. You are clearly extremely smart, have been able to adapt to various conditions after many years of playing professionally and whenever I have checked some of your heuristics/advices against my DB, it has always ended up positively. So an excellent or a bad year will do nothing to dissuade me from watching your videos.
However, knowing the recent results from a fresh new coach is somewhat valuable. It certainly doesn't give the whole picture about whether it is a coach worth watching but it is a good way to eliminate some coaches quickly: I have seen a few pros who became full time coaches when they stopped winning. Knowing a coach's recent results is always something I want to keep an eye on because it is an easy way to spot these coaches.
Be the first to add a comment