Betting Talk

Regression to the Mean Questions

donny2donny2 Senior Member
edited August 2018 in Sports Betting
I have a question about this and would like your guys opinions.



This is related to following cappers. Dr. H from what people mention is one of the best in mlb capping. He has been up a lot of units every single season. I noticed he got off in a fast start at the beginning... then cooled of just very little and well he's back to where he should be. He is up around 20 units already in mlb. Now being up 20 units where 1 bet is 1 unit should be a great season for any capper. But he averages much more than that per season. Now let say Dr. H at all star break is up like 35 units. If thats the case, is it true/false that there might be some regression to the mean? For example let say he has the same amount of plays in the 2nd half of the season to make it simple. Well it would be very hard for him to generate 35 units in the 2nd half of the season. To make it simple let say he averages 380 plays a season to make it simple. If after 190 plays, he is up 35 units, well shouldn't you feel that in the next 190 plays, it would be way too hard for him to generate the same amount of units? Meaning if it seems he average 35 units a season, well shouldn't you really expect there probably isn't going to be much profit since he's already up 35 units?


Now let say after half the season, Dr. H is up 5 units or breakeven or even down 5 or 10 units. If thats the case, wouldn't it make sense betting on all his plays on the 2nd half of the season makes much much more sense since variance got the best of him in the 1st half of the season? Thus you probably should even bet more than you normally did? My thoughts are if a guy who is very good is doing pretty bad in the beginning, assuming he doesn't do anything different, it seens like he will do much much better later on as there would be regression to the mean. Thus if he was down 5 units at the halfway mark of the season, i feel he would make a lot more units than in the 2nd half of the season if say he was up 25 units to make it simple.


I asked one capper about this and he tells me... the odds of that capper winning/losing should still be the same whether he was up 25 units or down 5 units for the 1st half of the season. I guess you can say its like roulette, it doesn't matter if it hit red 10 times... the odds of it being red the next one is still whatever it is etc. I mean if a capper goes on a very hot streak, i feel like tailing him after he's up a lot recently might be not a good idea. But if he's on a really long cold streak, well he should eventually hit his stride etc.

Comments

  • modeledthatmodeledthat Senior Member
    edited June 2017
    generally someone with a decent sample will evenly distribute their ev. it's hard to imagine someone falling flat or just completely losing their edge in the middle of a season. outside of that it's just usual variance. this thread is pretty aids
  • duritodurito Senior Member
    edited June 2017
    this thread is pretty aids

    this
  • FatCatFatCat Member
    edited June 2017
    Dr H has been runnin good on the side bets - especially the dogs. But I've been betting a little less than usual on his totals, which haven't been as good. The overs pick on the dodgers brewers on June 2 seemed crazy and I avoided that one. Still a big fan of his sides picks and I'll be sticking tight.
  • GoatsGoats Head Moderator
    edited June 2017
    You need to look up the definition of "regression to the mean."
  • RonbetsRonbets Senior Member
    edited June 2017
    Goats wrote: »
    You need to look up the definition of "regression to the mean."

    Why the effort? Just ask Siri and pose the question on multiple boards.
  • donny2donny2 Senior Member
    edited August 2018
    Okay what are your guys thoughts on this. Let say in the 1st half of the season, whether its an mlb team or nba team, let say they are down a lot of units like the kansas city royals. Or let say its an nba team that is like 8-37 and their ATS is like 12-33 halfway into the break. Or its a team thats like 33-12 ATS halfway through the break. In these situations, thoughts on betting on that team the rest of the season thats 12-33 ATS or fading the team that is 33-12 ATS? Because my thoughts are it would be very hard for that team 12-33 ATS to continue to lose that much ATS right? I would figure at worst, they go 50 percent ATS the rest of the season? The same can be said of the team that went 33-12 ATS? Surely they won't finish 60-22 ATS ... well it would be very hard right?
  • cshcsh Senior Member
    edited August 2018
    donny2 wrote: »
    Okay what are your guys thoughts on this. Let say in the 1st half of the season, whether its an mlb team or nba team, let say they are down a lot of units like the kansas city royals. Or let say its an nba team that is like 8-37 and their ATS is like 12-33 halfway into the break. Or its a team thats like 33-12 ATS halfway through the break. In these situations, thoughts on betting on that team the rest of the season thats 12-33 ATS or fading the team that is 33-12 ATS? Because my thoughts are it would be very hard for that team 12-33 ATS to continue to lose that much ATS right? I would figure at worst, they go 50 percent ATS the rest of the season? The same can be said of the team that went 33-12 ATS? Surely they won't finish 60-22 ATS ... well it would be very hard right?

    Every instance of a team covering a spread, winning a game, whatever you are betting on is an individual circumstance. Their frequency of doing the same thing in previous games has no bearing on future outcomes.

    Edit: individual circumstance assuming you are not discussing two correlated outcomes like spread & ML
  • RonbetsRonbets Senior Member
    edited August 2018
    I subscribe to something Mr. Hanahan/Hannaway(sp) taught us in Physics class. Gallileo's "law of inertia". Also Gordon Gekko, aka Michael Douglas would say, "trend is your friend".
  • jets96jets96 Senior Member
    edited August 2018
    csh , totally agree .
  • R40R40 Senior Member
    edited August 2018
    donny2 wrote: »
    Okay what are your guys thoughts on this. Let say in the 1st half of the season, whether its an mlb team or nba team, let say they are down a lot of units like the kansas city royals. Or let say its an nba team that is like 8-37 and their ATS is like 12-33 halfway into the break. Or its a team thats like 33-12 ATS halfway through the break. In these situations, thoughts on betting on that team the rest of the season thats 12-33 ATS or fading the team that is 33-12 ATS? Because my thoughts are it would be very hard for that team 12-33 ATS to continue to lose that much ATS right? I would figure at worst, they go 50 percent ATS the rest of the season? The same can be said of the team that went 33-12 ATS? Surely they won't finish 60-22 ATS ... well it would be very hard right?

    Yes, it will revert but the problem is that it is not likely to revert over the next 100 games. It will revert over many hundred games and not be profitable. The problem is that the team has been extremely unlucky most likely and the stat guys are good at figuring that out so the line does not adjust or they have been much worse/better than expected and people adjust to the higher/lower level of play so there is not much there for betting.

    There are opportunities though with these kind of streaks where a team has played much better or worse than expected when individual games have value as the market over-reacts. or the bad stats cause the modelers to underestimate a team and have a hard time catching up to a team's improvement. That happens quite a bit but not a situation for blind betting.
  • jets96jets96 Senior Member
    edited August 2018
    r40 , i wouldn't call myself a modeler at all ,but my model is set up in way to make sure i don't miss out on recent games ,it takes into account games from day one and as recent as yesterday ,I think if anyone has a model, recent play is incorporated, how one weighs those games are very important.
    It gets harder come sept , am sure that's why the drh's of the world stop with their services so soon.
  • R40R40 Senior Member
    edited August 2018
    jets96 wrote: »
    r40 , i wouldn't call myself a modeler at all ,but my model is set up in way to make sure i don't miss out on recent games ,it takes into account games from day one and as recent as yesterday ,I think if anyone has a model, recent play is incorporated, how one weighs those games are very important.
    It gets harder come sept , am sure that's why the drh's of the world stop with their services so soon.

    Yes, but a model is highly dependent on stats. With small sample sizes always being an issue, it is tough for a model to get the right number when a team has a turnaround and the numbers will lag. That is where I see models being particularly weak and often good numbers are created when the market is dominated by math models.
  • jets96jets96 Senior Member
    edited August 2018
    R40 , why do you think its hard to get the right number when a team has a turnaround, I think the complete opposite, it's really easy.
    FOR EXAMPLE , one part the model asks what a team did its last 10 games , last 5 games , last series , am saying this in the simplest terms, many different offshoots to that, but you know what am saying.
    For the most part , lines are derived from that days pitcher , not so much about the TEAM.
  • R40R40 Senior Member
    edited August 2018
    jets96 wrote: »
    R40 , why do you think its hard to get the right number when a team has a turnaround, I think the complete opposite, it's really easy.
    FOR EXAMPLE , one part the model asks what a team did its last 10 games , last 5 games , last series , am saying this in the simplest terms, many different offshoots to that, but you know what am saying.
    For the most part , lines are derived from that days pitcher , not so much about the TEAM.

    I'm not referring to baseball at all which there is a lot of information to base a model on. More referring to football and basketball which is much more difficult to get a handle on and complete turnarounds are not uncommon.
  • donny2donny2 Senior Member
    edited August 2018
    Well my thoughts are this. If its football, its harder because there are much fewer games and thus smaller sample. Thus if some team goes 1-7 ATS, well it wouldn't be surprising if they finished 3-13 ATS. But in say nba and mlb, wouldn't that be a bit different since bigger sample? I know you say you don't know how long it take to revert etc.
  • R40R40 Senior Member
    edited August 2018
    donny2 wrote: »
    Well my thoughts are this. If its football, its harder because there are much fewer games and thus smaller sample. Thus if some team goes 1-7 ATS, well it wouldn't be surprising if they finished 3-13 ATS. But in say nba and mlb, wouldn't that be a bit different since bigger sample? I know you say you don't know how long it take to revert etc.

    No. There is probably some value on the team for a few games. But to take advantage (if any existed) you'd need to know when to get off, and if you are betting with that strategy, it is impossible. They are not any more likely to beat the spread just because they haven't. They will inevitably go 13-3 ATS but not likely in the near future.
Sign In or Register to comment.