why the followers are more likely to lose

[Data]

Uh.......... in backtesting, yeah, but that's not a meaningful concept when going forward since there's no defined finite future bet bin to be drawing from.

Say, a season. You actually have to define the bet bin otherwise you cannot define the winrate. In conditions we are dealing with, there is no future winrate, there is only near future winrate. The winrate varies, the history matters.

[stuckinnj]

This is the (valid) "my lane is always slower" theory of traffic jams. You notice the other lane moving and move into it only after a good bit of the slack is taken up, and then fail to benefit from slack in your old lane. I'm with you Data.

[Thremp]

Data,

In earnest, you are being incredibly retarded to not understand Cowley's point. That or deliberately deceitful.

[tomcowley]

Say, a season. You actually have to define the bet bin otherwise you cannot define the winrate. In conditions we are dealing with, there is no future winrate, there is only near future winrate. The winrate varies, the history matters.

Then you just have a form of fallacy via selection bias. Even though the capper's EV, before the season starts, is 54%*#plays, when you wait to jump on until some statistical test is passed, he's going to pass it way, way, way more often when he's run way above expectation on a relatively small number of bets than when he's run at expectation for a larger number of bets. When you jump on, at any point, your expectation is 54%*#remaining plays.

Compare a couple of basic strategies- "believe" the capper is 54% before the season and follow for the whole season. Your EV/play going forward is based on 54%. Wait for the capper to "prove it" this season before jumping on. Your EV/play going forward is also based on 54%. What you're comparing is apples and oranges- you're comparing the results with making him "prove it" to the results of following for the whole season WHEN HE WOULD HAVE PROVED IT. That cuts out all the times you would have been following him and he would have lost early and you never would have jumped on. It's not a viable strategy to wait and see if he runs hot, then go back in time and follow him from the get-go.

[Data]

Then you just have a form of fallacy via selection bias. Even though the capper's EV, before the season starts, is 54%*#plays, when you wait to jump on until some statistical test is passed, he's going to pass it way, way, way more often when he's run way above expectation on a relatively small number of bets than when he's run at expectation for a larger number of bets. When you jump on, at any point, your expectation is 54%*#remaining plays.

Please define what do you mean by providing a certain number (i.e. 54%) going forward to the near future. Make sure that capper's edge may vary by game.

[tomcowley]

Read the stanford encyclopedia of philosophy entry on probability if you want to get into the interpretations (I'm personally a fan of propensity as the proper application in most real-world cases). The definition you're wanting to use, "will hit exactly x% in the next fixed number of games" is obviously not applicable to any sports capper (well, I guess the guy who always bets both sides of every game at the same time, but he's not interesting), and you need something that pathological to use your sampling-without-replacement argument.

[Data]

Going forward, the series of all bets and any subseries of those bets (chosen before the results of the bets being chosen are known) have to converge to the same long-term winrate as long as all the bets are independent.

So far I showed that the bets are not independent. Now, I am getting at another issue, there is no such thing as long-term winrate for a capper. There is no such quality for a capper to beat the lines at certain percentage, it is not up to him alone. What he does is finding betable lines. His edge understandably varies per bet. The lines are getting sharper, the betting opportunities dry up, the winrate goes down. You cannot look at series of bets as a series of Bernoulli trials because these processes are too different.

[Data]

Going forward, the series of all bets and any subseries of those bets (chosen before the results of the bets being chosen are known) have to converge to the same long-term winrate as long as all the bets are independent.

Part 3. While there is no such a thing as a winrate, the edge on a given bet is real and known to a capper before he makes the bet. Therefore, there are "subseries of those bets (chosen before the results of the bets being chosen are known)" that have more or less than average edge.

[Thremp]

Then why did you deliberately deceive everyone by citing a winrate. Elaborate trolling IMO.

[tomcowley]

Seriously? Your entire argument is that followers have a lower EV/bet than a winning capper because a capper's edge always declines over time, and they miss the early (best) bets, and you went through all of that to get there?

[tomcowley]

Part 3. While there is no such a thing as a winrate, the edge on a given bet is real and known to a capper before he makes the bet. Therefore, there are "subseries of those bets (chosen before the results of the bets being chosen are known)" that have more or less than average edge.

You're the one who stipulated a 54% capper and a winrate.

Let's simulate a "leader" (L) that is 54% and starts with 0-0 record, and a follower (F) who does not know the Ls rate.

GFY.

[Data]

Data,

In earnest, you are being incredibly retarded to not understand Cowley's point. That or deliberately deceitful.

Thremp, I have no problem with you believing that you are so much smarter than me. I actually hope that you are right, or, at least, not wrong by far. Because if that is so, there is a chance that we will see you contributing. I am unsure if you are trying engage me in conversation or just trying to use me to show off something to somebody. At this point I look at you as a kid who has very large understanding gaps to fill. You seem to have some potential but that is all. At this point of your development I have no interest conversing with you as you have provided no evidence that you have something to offer in conversation that I could be interested in. Again, I don't mind you running around and barking loud, it's kinda cute for in a puppy way. Try learning human language and putting together more than two sentences every once in a while. Don't be so afraid to say something stupid. With saying so little you mange to do this anyway. There is no shame to be wrong. You just gotta improve and that is something I wish you and the others. Talk to you later, I hope.

[Data]

If the EV of every bet is independent, long-term timing of noise will just change the results to EV/play*#plays*(%plays played) instead of EV/play*#plays.

Here you introduced approaching this problem as Bernoulli trials. I merely followed looking where this gets us. Perhaps I should not have. I found that this model is too far from reality. Still, I think it was worth exploring as this involves thinking about and discussing some important things. It is one of my favorite subject to reflect upon applicability of certain methods and theories to given problems. Before I GFM I want to thank you for an interesting discussion.

[dimon]

here is a scenario, that I used this year...no hard numbers like you all use, but...capper is solid, and wins at 54-56% rate, and I belive that he can keep at that rate...now the way I look at it, is similar to old fella Buffet - "by low, sell high"...capper goes 23-7 on the series of bets...I start fade the guy, no matter what...he goes on a loosing streak...I pick him up...just like a stock that good and you know it, but no one likes

[tomcowley]

Here you introduced approaching this problem as Bernoulli trials. I merely followed looking where this gets us. Perhaps I should not have. I found that this model is too far from reality. Still, I think it was worth exploring as this involves thinking about and discussing some important things. It is one of my favorite subject to reflect upon applicability of certain methods and theories to given problems. Before I GFM I want to thank you for an interesting discussion.

The problem is that expecting a capper's ev to oscillate like that is fairly silly on short timescales- it's certainly possible seasonally (crush the first couple of weeks because you did a lot of homework, flip the rest, rinse, repeat), but expecting multiple oscillations in that pattern in a season is pretty odd.. and far more likely to be noise IMO. You'd have to be really skilled in picking up dynamic edges combined with being really slow to see that they didn't exist anymore (obviously you're dealing with timescales where bets won/lost isn't close to converging to any decent confidence levels) to get patterns like that. I would expect that far, far more followers lose because they are following something that never was +EV and just ran hot, or because they were following something that was +EV, but isn't anymore, and probably never will be again anytime soon.

[Salamander]

There is a box filled with 54 blue balls and 46 red balls. I can bet $100 to win $100 on blue. After a ball is chosen it is put back in the box and the box is shaken up well.

Imagine there are 100,000 such boxes.

100,000 "cappers" stand at the ready in front of these boxes. They now all choose 10 balls. Let's say that 15% of them pick at least 8 of 10 blue balls. These are now the handicappers to be followed now because they have COME TO OUR ATTENTION as having done so well.

Of this new sample of 15,000 "cappers", they will again choose 10 balls. Again, 15% will continue well and choose at least 8 of 10 blue balls while a similar (but smaller) amount of "cappers" will choose only 2 or fewer 10 blue balls.

But you the follower aren't getting "on" or "off" anything. The 54% cappers are just continuing in their 54% ways. Whether you followed one that is continuing on a good streak or has gotten cold in the short term is completely random.

[Thremp]

Thremp, I have no problem with you believing that you are so much smarter than me. I actually hope that you are right, or, at least, not wrong by far. Because if that is so, there is a chance that we will see you contributing. I am unsure if you are trying engage me in conversation or just trying to use me to show off something to somebody. At this point I look at you as a kid who has very large understanding gaps to fill. You seem to have some potential but that is all. At this point of your development I have no interest conversing with you as you have provided no evidence that you have something to offer in conversation that I could be interested in. Again, I don't mind you running around and barking loud, it's kinda cute for in a puppy way. Try learning human language and putting together more than two sentences every once in a while. Don't be so afraid to say something stupid. With saying so little you mange to do this anyway. There is no shame to be wrong. You just gotta improve and that is something I wish you and the others. Talk to you later, I hope.

Why did you cite a winrate when you said you explicitly do not believe in that such a thing exists?

[Data]

Why did you cite a winrate when you said you explicitly do not believe in that such a thing exists?

It is a part of tomcowley's model of the process we are studying. He introduced a model that is otherwise widely used. That model called for results contradicting with an existing (I am not claiming the theory that I stole as mine) theory and with anecdotal evidence. Since this model is so widely used it was worth exploring and looking closer at to see how it "fits" the process we we are studying. Apparently, it does not.

[Thremp]

It is a part of tomcowley's model of the process we are studying. He introduced a model that is otherwise widely used. That model called for results contradicting with an existing (I am not claiming the theory that I stole as mine) theory and with anecdotal evidence. Since this model is so widely used it was worth exploring and looking closer at to see how it "fits" the process we we are studying. Apparently, it does not.

Wow.

[PRC]

Thremp, just saying "wow" doesn't make you seem smart. While tomcowley and data clearly disagree, at least they are making intelligent points. Try saying something useful for a change.

[Data]

The problem is that expecting a capper's ev to oscillate like that is fairly silly on short timescales- it's certainly possible seasonally (crush the first couple of weeks because you did a lot of homework, flip the rest, rinse, repeat), but expecting multiple oscillations in that pattern in a season is pretty odd.. and far more likely to be noise IMO.

Every capper's play has a different EV. It is silly not see that EV over several plays will ocsillate. Again, there is no such thing as capper's EV. The EV depends on capper's skills AND the lines he is dealt. The plays are without replacement. The plays where the capper has highest +EV are without replacement. If he runs hot he likely is hitting those plays with the highest +EV. Now, the followers joins. The capper's average EV/play at this point is likely lower than it was before the season. Let's see what if it is still positive. If the follower stays with the capper he will have the same still positive results albeit likely of a lesser degree. If he abandons the capper after a series of losses, misses a hot streak, and then rejoins the capper, then his EV is lower than the capper's because he likely missed games with higher +EVs.

[Thremp]

Every capper's play has a different EV. It is silly not see that EV over several plays will ocsillate. Again, there is no such thing as capper's EV. The EV depends on capper's skills AND the lines he is dealt. The plays are without replacement. The plays where the capper has highest +EV are without replacement. If he runs hot he likely is hitting those plays with the highest +EV. Now, the followers joins. The capper's average EV/play at this point is likely lower than it was before the season. Let's see what if it is still positive. If the follower stays with the capper he will have the same still positive results albeit likely of a lesser degree. If he abandons the capper after a series of losses, misses a hot streak, and then rejoins the capper, then his EV is lower than the capper's because he likely missed games with higher +EVs.

o rly

[Data]

o rly

[Thremp]

What do you mean by "run hot"?

[beckluna]

i always seem to lose when i follow someone

[wrongturn]

It is a non-issue. If one has to follow others because he couldn't win by himself, he is no more likely, or less likely for that matter, to lose by following.

[saratoga1927]

It is a non-issue. If one has to follow others because he couldn't win by himself, he is no more likely, or less likely for that matter, to lose by following.

The thread is just hot air from the git go!

[tomcowley]

Every capper's play has a different EV. It is silly not see that EV over several plays will ocsillate. Again, there is no such thing as capper's EV. The EV depends on capper's skills AND the lines he is dealt. The plays are without replacement. The plays where the capper has highest +EV are without replacement. If he runs hot he likely is hitting those plays with the highest +EV. Now, the followers joins. The capper's average EV/play at this point is likely lower than it was before the season.

Why?

[beckluna]

hit 35%

[Data]

What do you mean by "run hot"?

To make an observer to notice.

[Data]

Why?

It is a logical conclusion from the observations below. That is why it followed them. The season is short, the betting opportunities with high +EV are scarce.

The plays where the capper has highest +EV are without replacement. If he runs hot he likely is hitting those plays with the highest +EV.

[Thremp]

To make an observer to notice.

Well if this isn't the most chickenshit response I've read in the last 2 posts.

[Data]

Well if this isn't the most chickenshit response I've read in the last 2 posts.

But then are not I at least smart to be shit scared? You sound like a mean real badass. You must be one, I do not want to mess with you.

[Maverick22]

Thremp, I have no problem with you believing that you are so much smarter than me. I actually hope that you are right, or, at least, not wrong by far. Because if that is so, there is a chance that we will see you contributing. I am unsure if you are trying engage me in conversation or just trying to use me to show off something to somebody. At this point I look at you as a kid who has very large understanding gaps to fill. You seem to have some potential but that is all. At this point of your development I have no interest conversing with you as you have provided no evidence that you have something to offer in conversation that I could be interested in. Again, I don't mind you running around and barking loud, it's kinda cute for in a puppy way. Try learning human language and putting together more than two sentences every once in a while. Don't be so afraid to say something stupid. With saying so little you mange to do this anyway. There is no shame to be wrong. You just gotta improve and that is something I wish you and the others. Talk to you later, I hope.

wow. Data just OWNED Thremp.

Data is the sheriff of the Think Tank!!! Laying down law!

[tomcowley]

It is a logical conclusion from the observations below. That is why it followed them. The season is short, the betting opportunities with high +EV are scarce.

This still looks like the card-counting fallacy to me. Before the season, the capper thinks that the odds of his second play being 60%+ are x. His first play of the season is 60%+. Is x higher or lower, and why? What if his first play of the season is 55%? Is x higher or lower, and why? (assume he won't skip any games where he thinks he has a sufficient edge, and he expects to have nonzero numbers of plays everywhere from 54% to 60%+).