We all project. We all convert.
April 19th, 2010 by Robert Dixon in General Guidance, TheoreticalThis is a claim I've made a few times in the comments and before any more veins pop in any more foreheads I am going to clarify what I mean by this.
Let's say you are playing poker. It is down to you and one other person on the river. The pot has $800 in it and he bets $100. If you call that $100 you are "saying" that you think you are 10% or greater to have the better hand. Whether you calculate pot odds or not. Whether you try to read your opponent or not. Presumably, your goal is to win money and to that end your actions are claiming a 10% or better shot at winning.
The same is true in a FBB auction. When you buy a player you are making a statement about the stats that player will produce and what that production is worth. Your idea about his projected stats might be a somewhat vague one based on a lot of small factors from which you choose those you think are most pertinent. Your converter might be years of experience and faith that this allows you to accurately translate that value. Regardless, you are still projecting and converting. I'd go so far as to say every time someone else buys a player in an auction you are doing the same thing but this time you are synthetically selling the player by refusing to bid one more.
If you truly aren't projecting or converting on any level I think you are picking your players at random and flipping a coin to decide whether or not to call on the river.
Loading ...
I think I addressed this point already as it relates toward building a model or using my "gut algorithm". Sure – at some level the factors I consider are some kind of a projection insofar as they're a basis for picking a player. But I do not do it consciously, i.e., I do not choose specific best-guess numbers intentionally as my basis for buying a player. I simply let my instincts take over when that player's name comes up. I leave the idea of his production purposely vague in my conscious mind, while my instincts to bid or stop are informed by "feel". I don't stop bidding because of any kind of conscious and specific projection has been exceeded. I stop because the player no longer feels like a good bargain at the going rate.
So my algorithm projects in the loose sense that you describe above, but I'm not conscious of any numbers involved except vaguely. If I were playing poker, I would consciously calculate the pot odds there, obviously.
I think that's Robert's point exactly. That even though you don't do it consciously, your actions still have mathematical statements necessarily attached. Just like the poker player who says "I just called, I didnt do any math there, but It felt right" is mathematically tantamount to saying (in the above example) "I thought I was at least 10% right to have the best hand."
Bidding $41 for Arod is making a statement that you think he will likely produce output at or above that price. That's really all Robert's saying here.
why would you bother to consciously calculate pot odds when you are making only rough estimates of what cards your opponent is holding? Isn't this exactly the same as making rough projections about a player but then converting it into precise dollar values?
Not at all, Chris. The pot odds are a fact. There is no speculation there. Why would I not inform myself of all of the pertinent facts before making a decision? For me not to calculate the pot odds would be like me not knowing the entry fee for the league, or not knowing how many pitchers I need to start each week.
You're right. But you can't do anything with the pot odds unless you have some projections for your opponent's likely hand ranges. Otherwise you have nothing to compare them to.
Once I make some estimations of my opponent's likelihood of each holding, I can then perfectly compute how my holding performs against his holding (even if there are more cards to come) and decide what to do based on the pot odds.
We can do this because we have a perfect mathematical basis to come up with a precise number once we have all the inputs (even if some are best guesses). This is exactly analogous to a perfect pricing model, and this is a good thing to have. Your argument has been analogous to saying it doesn't help me to know how to perfectly translate my projections/hand ranges into a mathematical solution, and that is Chris Pikula's point.
It would even be reasonable to project more than one value based on preformance. If he has a bad year his stats are… if he has a good year his stats are… his average year is… then you can have an estimate of fair value, and a high/low range to go with it.
I don't think it's really the same kind of analysis going on when you imagine what a player can do on the field as when you imagine what cards someone could have.
In one you're predicting the future, in the other you're predicting the past. (What cards must he have had to do x, y and z). For tasks like the former, models do a poor job in my opinion. In tasks like the latter, they're very good.
We're much better at solving the mysteries of the past through analytical thinking (which is what the model is a tool for) than predicting the future.
Can you substantiate that claim at all?
i will in a separate post.
Other than this claim, would you also concede the point? Or do you also object on other grounds?
See the next post.