I know that chasing is super controversial. I had the good fortune of corresponding with Justin7 and he pointed me in the path that the pros use. While his logic makes perfect sense to me, I am still fascinated by the corrective quality of chasing.
I have spent the last couple of years studying different systems and scenarios. I recently studied a scenario that I believe is a game changer. First I will explain the main argument, the scenario, and then the theory (or perhaps it would be more appropriate to call it a hypothesis at this point).
These sites can be vicious so let me just state that I do not think I am a genius, or any such foolishness. I am just an average guy who finds it entertaining to think about this stuff. These are my thoughts, nothing more, nothing less.
Whether you call it chasing, the progressive system, or the Martingale System the main argument is the same.
The main argument:
“Progressive betting is not a legitimate money management device, but many fraudsters try sell systems using forms of progressive betting... ...no system can turn a negative EV scenario into a positive one.”
- “Conquering Risk” by Elihu D Feustel and George S Howard pg. 43
The main foundation for this argument is that the results are random, such as a coin toss, spin of the roulette wheel, or the roll of dice at the craps table. The theory I am proposing is that sportsbetting is not random. It may appear to be but it is anything but random. The spread is determined by public opinion. In order to maintain a 50/50 split on underdogs and favorite it is necassary for public opinion to switch from undervalued to overvalued, and back. I call this variance the "Pendulum Cycle".
I first discovered it by studying the 2007 Patriots season. This season was an abberation. They beat the spread on the first 8 games. They were clearly undervalued by the public. As each game progressed the point spreads increased in size. This was not completely linear, but close. The Bengals and the Colts upset the pattern because they were considered to be stronger opponents then the others.
The Patriots lost ATS in the Colts game by 1 point. I suggest that if they had beat the spread in that game that the Bills game that followed would have had a spread over -20 rather then -16. Instead this did not happen until the Eagles at -24. This was the crux event. Had you stuck with the underdogs for the rest of the season you only would have lost once with the Steelers.
This was clearly a bubble phenomenon... ...just like the stock market. What were the odds of the Patriots actually being able to beat a -24 spread against the Eagles in that game. 50/50??? Hell no!. I am not a statistician, but if I had to guess I would say they had less then a 20% chance of covering the spread in that game. Did this streak end with the Eagles? No! With only one exception, the underdogs prevailed for the rest of the season.
This season was an abberation. It is not the norm. Typically speaking, most teams should go 3 to 4 games max before losing ATS. Anything beyond this is abberant behavior and a backlash isn't just likely, it's inevitable.
With each game that the Patriots beat the spread, the next game would have a spread that was more difficult to overcome. This does not fit the negative EV model presented in the main argument against systems. With each successful cover the next game falls further and further from 50%... ...which means that the underdogs chances in these games are increasing well over 50%. This is creating a positive EV scenario if you use a chase system. You can't be certain when the streak will break, just that the odds of it continuing are not random. The gambler's fallacy actually applies in this scenario.
There are two approaches that you can use this information with if you are into chase systems. When you find an abberation you can ride a favorite until they lose ATS and then switch to their opponents. A chase system would have served you well in the case of the Patriots. This was an extreme case, but far from the only one. I used a similar approach on the Lions and 49ers for 2011. I played them for the first 8 and switched to their opponents for the second. The results were killer. I did the same thing with the Packers. Though it came out in positive territory, it would have been better to stick with the Packers for the entire season.
This idea is in its infancy. I believe with enough study I will eventually be able to chart this information and literally be able to use technical analysis to pick teams. The problem is that there are two teams that must be accounted for. They are each in separate cycles. I have not yet figured out how to chart this in a manner that is a usable forecasting tool, but the thought is intriguing.
It is too simple to just say "fade the public". They are right 50% of the time. The entire point spread system depends on this fact. It is the frequency between them being right vs wrong that is the key factor. Some teams remain in short cycles while others go on abberant streaks. Both pieces of information are useful. You can chase a team in a short cycle, or you can wait to buck the trend of a long streak.
I humbly present the idea that results ATS are not random but cyclical. This negates the negative EV scenario argument for each successful game ATS increases the EV of the next game for the opponent, and vice versa. (aka not random) By understanding the cycles it is theoretically possible that chase systems are a legitmate tool.
I have been backtesting an idea to test the random theory. I have a very deep chase system, many would consider impractical at 7 levels, but those are my perameters. It pulls an average return of 10% to 15%. Not super exciting until you realize how simplistic my filter was for choosing teams.
In order to make things random I went to NFL.com and chose the first Sunday game listed in the regular season for all 17 weeks. I chose the last Sunday game because it would be played at 8:00 pm. This gives enough time for the chaser to see the results of the first game and place a wager for the late game. I then chose the Monday night game. If there was more than one I took the first. (There was at least one scenario where there was more than one Monday nighter, but it was rare) This gave me 50 games for the regular season. There is no Monday night game in week 17. I left out the Thursday games and the occasional Saturday games. I did this simply because 50 is a nice round number. I chose to bet strictly on the favorites using my seven tier chase. I then used the same games for the underdogs. I have done this for 5 seasons; 2007-2011. I hate to sound like Jon Morrison but the success rate is 100%. This is very time consuming, but I want to take this all the way back to 1993. So far I have only made it to tier 7 once. Since I have used both favorites and underdogs this is a total of 500 trials. If I make it to 1993 without a failure it would hardly be conclusive evidence, but it should be enough to convince people that maybe this chasing thing warrants a little investigation.
I hope somebody finds this as interesting as I do. May you always win more than you lose.