There obviously isn't an "absolute" answer to when something goes from "short term" to "long term". Some might feel comfortable following a handicapper after seeing 100 plays from them. Others will want to see 1000. Is one season enough, or do you need 3 or 5 seasons? There aren't definitive answers to those questions, however what we're trying to limit the line of thought that someone's success (or lack thereof) in their last 5, 10, or 15 plays has any meaning when compared to a much larger sample size for that person.
I get the feeling that many (or at least some) would feel more confident tonight following a lifetime 50% handicapper that's gone 10-1 over their last 11 plays vs. a lifetime 55% handicapper that's gone 1-10 over their last 11 plays. The "hot and cold" thread was, at least in my opinion, intended to say that statistics will tell you that the 55% handicapper will have several very good runs and several very bad runs over the course of a lifetime, simply due to variance. At the same time, a losing better will certainly have some extremely good runs if they have enough volume, simply due to variance. But at the end of the day, the lifetime record is a much better indicator of future success that their record over the last "x" number of plays. (and of course guys can gain/lose an edge over time, so that's just another thing to throw in there when figuring out who to follow, however someone doesn't gain/lose their edge over 5 or 15 or even 50 plays unless something drastic has occurred).
Excellent post. You have stated your position very well, and alot more informative/instructive than "over time". Maybe a stickey to reference.
Now, IF my choice today is to follow the 50% vs. the 55% historical posters in your example, I am likely to follow the 10-1. That IS NOT because I respect/disrespect one over the other, just that I am making a choice today, not for the next week or month or year. AND, also, that 10-1 or 1-10 GUARANTEE absolutely nothing. Just as has been stated, the streak is the past, so is the 50% or 55%.
Also, I do admire that You, Goat's, Ed and many others chiming in, who follow your methodology religiously/relentlessly/strictly because I persionally believe that discipline in following ones chosen methodology is far more important to succss than any random string of picks.
Now, IF my choice today is to follow the 50% vs. the 55% historical posters in your example, I am likely to follow the 10-1. That IS NOT because I respect/disrespect one over the other, just that I am making a choice today, not for the next week or month or year. AND, also, that 10-1 or 1-10 GUARANTEE absolutely nothing. Just as has been stated, the streak is the past, so is the 50% or 55%.
The point is which option is more likely to continue to be profitable? The 50% handicapper who is "hot" or the 55% handicapper who went 2-8 this week.
I think everyone gets the point, except for dragonswirl.
If a handicapper doesnt have an edge, following any play of theirs is -EV, and if a handicapper does have an edge, following any play of theirs is +EV. This holds true regardless of any current good/bad streak, because the streak has no bearing on the result of the next play.
Following a capper with no edge because of a 10-1 streak (or any other reason for that matter) is -EV, and that is the main point Goats was trying to make when starting this thread.
It would be like a casino closing the craps table (where they have a predetermined edge) because someone was "hot", you just would never see that happen.
You could go through the entire RAS archive, and look for plays after a bad run, pick any random run you want, lets use 1-4 so the sample size will be big enough, and I would bet that the 6th play following a 1-4 run is similar to our long term win rate, and similar to the 6th play after a 4-1 run as well.
I think everyone gets the point, except for dragonswirl.
It would be like a casino closing the craps table (where they have a predetermined edge) because someone was "hot", you just would never see that happen.
Ed,
Maybe the more appropriate craps metaphor is IF one is a don't bettor and has a choice of two tables...One where there is whooping and hollerin' and proverbially "hot", passes to beat the band...or a quiet, dour table with glum faces and point, seven-out, point, seven-out...
IF your choice is the "Happy Table", God Love ya (been there, done that, and got more than enough tee-shirts...don't do it any more, no matter what the math says). Neither condition is either permanent, nor does is currently exist. I'll still take the glum group, with a plan to exit when/if they get "happy". Call it experience.
As to +/-EV, we had this discussion, part of EV is MY or YOUR expectation. I respect that You wouldn't make the same choice.
"You could go through the entire RAS archive, and look for plays after a bad run, pick any random run you want, lets use 1-4 so the sample size will be big enough, and I would bet that the 6th play following a 1-4 run is similar to our long term win rate, and similar to the 6th play after a 4-1 run as well."
I would expect the same as well, Just as I would expect the same at 1-2/2-1 which became 1-3/3-1 before becoming 1-4/4-1.
And I am actually going through RAS picks performance at Final line, probably shouldn't waste my time, as I doubt you would let someone who doesn't "Get It" to subscribe anyway. ;-)
At what point do you begin to discount the back portion of LT records? Obviously handicappers can improve or the can fall behind, so at what point do we no longer consider the handicapper as having a 55% edge who was 75% from 2002-2004 and 40% since? Judgement call? 5yr win percentage? Any opinions? Is the full LT record even relevent with someone like that?
Good question, something as dramatic as your example would have to be scrutinized, but in general I would say if you were looking at a 1000 game sample size over five years, you would give slightly more weight to the most recent years. The biggest things to look for is what if anything has changed with the handicapper (this can sometimes be difficult if the person is not forthcoming) and what has changed in the market. Is the market more efficient, have other handicappers who won previously now having a harder time to win, etc.
At what point do you begin to discount the back portion of LT records? Obviously handicappers can improve or the can fall behind, so at what point do we no longer consider the handicapper as having a 55% edge who was 75% from 2002-2004 and 40% since? Judgement call? 5yr win percentage? Any opinions? Is the full LT record even relevent with someone like that?
It's impossible to come up with an exact answer IMO and is definitely at least part judgment call as you suggested. It's also helpful to have more than just a capper's record to go on.
Using StevieY as an example, if you read his recent posts, he discussed how his records declined in the late 2000's, so he put a lot of work into his computer models, came up with what he felt would be significant improvements, and then won almost 50u in CBB last year while hitting approximately 59%. That's been followed up by him turning a profit across all 6 sports thus far in the 2010-2011 seasons here at BT. Based on his past records alone, it would have been much tougher to be as confident going into the most recent seasons, but the added info he shared (and of course my personally knowing him to be as intelligent an hard-working as it gets for the past decade) certainly adds a lot to the decision-making process of whether to follow or not.
In general, I would put more weight on recent years due to the market evolving, but I wouldn't completely discount records from further back.
Again, these are just my opinions. I have written many times that deciding whom to follow is as much art as it is science. If there was a formula we could apply, it would be far easier, but of course there's not... and we should always remember the caveat: "past results are no guarantee of future success."
Comments
Excellent post. You have stated your position very well, and alot more informative/instructive than "over time". Maybe a stickey to reference.
Now, IF my choice today is to follow the 50% vs. the 55% historical posters in your example, I am likely to follow the 10-1. That IS NOT because I respect/disrespect one over the other, just that I am making a choice today, not for the next week or month or year. AND, also, that 10-1 or 1-10 GUARANTEE absolutely nothing. Just as has been stated, the streak is the past, so is the 50% or 55%.
Also, I do admire that You, Goat's, Ed and many others chiming in, who follow your methodology religiously/relentlessly/strictly because I persionally believe that discipline in following ones chosen methodology is far more important to succss than any random string of picks.
Regards,
The point is which option is more likely to continue to be profitable? The 50% handicapper who is "hot" or the 55% handicapper who went 2-8 this week.
If a handicapper doesnt have an edge, following any play of theirs is -EV, and if a handicapper does have an edge, following any play of theirs is +EV. This holds true regardless of any current good/bad streak, because the streak has no bearing on the result of the next play.
Following a capper with no edge because of a 10-1 streak (or any other reason for that matter) is -EV, and that is the main point Goats was trying to make when starting this thread.
It would be like a casino closing the craps table (where they have a predetermined edge) because someone was "hot", you just would never see that happen.
You could go through the entire RAS archive, and look for plays after a bad run, pick any random run you want, lets use 1-4 so the sample size will be big enough, and I would bet that the 6th play following a 1-4 run is similar to our long term win rate, and similar to the 6th play after a 4-1 run as well.
Ed,
Maybe the more appropriate craps metaphor is IF one is a don't bettor and has a choice of two tables...One where there is whooping and hollerin' and proverbially "hot", passes to beat the band...or a quiet, dour table with glum faces and point, seven-out, point, seven-out...
IF your choice is the "Happy Table", God Love ya (been there, done that, and got more than enough tee-shirts...don't do it any more, no matter what the math says). Neither condition is either permanent, nor does is currently exist. I'll still take the glum group, with a plan to exit when/if they get "happy". Call it experience.
As to +/-EV, we had this discussion, part of EV is MY or YOUR expectation. I respect that You wouldn't make the same choice.
"You could go through the entire RAS archive, and look for plays after a bad run, pick any random run you want, lets use 1-4 so the sample size will be big enough, and I would bet that the 6th play following a 1-4 run is similar to our long term win rate, and similar to the 6th play after a 4-1 run as well."
I would expect the same as well, Just as I would expect the same at 1-2/2-1 which became 1-3/3-1 before becoming 1-4/4-1.
And I am actually going through RAS picks performance at Final line, probably shouldn't waste my time, as I doubt you would let someone who doesn't "Get It" to subscribe anyway. ;-)
Regards,
It's impossible to come up with an exact answer IMO and is definitely at least part judgment call as you suggested. It's also helpful to have more than just a capper's record to go on.
Using StevieY as an example, if you read his recent posts, he discussed how his records declined in the late 2000's, so he put a lot of work into his computer models, came up with what he felt would be significant improvements, and then won almost 50u in CBB last year while hitting approximately 59%. That's been followed up by him turning a profit across all 6 sports thus far in the 2010-2011 seasons here at BT. Based on his past records alone, it would have been much tougher to be as confident going into the most recent seasons, but the added info he shared (and of course my personally knowing him to be as intelligent an hard-working as it gets for the past decade) certainly adds a lot to the decision-making process of whether to follow or not.
In general, I would put more weight on recent years due to the market evolving, but I wouldn't completely discount records from further back.
Again, these are just my opinions. I have written many times that deciding whom to follow is as much art as it is science. If there was a formula we could apply, it would be far easier, but of course there's not... and we should always remember the caveat: "past results are no guarantee of future success."