es330td
• Posts: 151
Joined: Mar 19, 2019
August 29th, 2020 at 1:13:43 PM permalink
Okay, I need to post several things. I ask anyone following this refrain from commenting until I finish.

The system is based on a modified Labouchere cancellation betting progression.

As discussed in other posts, there are two aspects to my strategy: picking games and bankroll management.

In reading Mike Shackleford's post about the Labouchere system I noticed that all the simulations were based on bets with an even money payoff and a roughly 50% chance of winning, red/black in roulette, pass line in craps, baccarat. A thought occurred to me, "How would this system work if the gambler was favored to win his bet?" I started looking around trying to find bets where the gambler is favored to win, even if the payoff is lower than 1:1. There aren't many, but they do exist. The first thing I tried was betting on two thirds in roulette. In this situation you have a 24/38 chance of winning 1 but a 14/38 chance of losing two. This win rate of 63% still has a fairly high chance of failure. I did some research and found out that in general sports bets of -250 or greater have a 75% win rate. The chance of seven consecutive losses is 0.006%, small enough to be very unlikely. Not impossible but highly improbable. I downloaded baseball historical outcomes and ran tests to confirm that selecting home favorites between -250 and -300 worked. Not quite 75% win rate but close.

The second problem is bet sizing. Losing a -250 bet makes the added bet pretty big. If your progression is 1 1 1 1 1, losing makes the progression 1 1 1 1 1 5. If your next wager needs to win 6, losing that makes the next bet 15. What I did is modify the system so that if I lost a game I would still add the loss to my progression but then I would average the values so instead of 1 1 1 1 1 5, it would be 1.66 1.66 1.66 1.66 1.66 1.66. The bets still increase like any other progressive betting system but more slowly. The thought is that it is okay to add more numbers to the progression because on average I am going to win roughly three times as often as I lose. It all seems complicated but when you have a spreadsheet to keep track of the amounts it is pretty easy. This would never work live at a table game but when you only make one wager per day it is simple.
es330td
• Posts: 151
Joined: Mar 19, 2019
August 29th, 2020 at 1:15:45 PM permalink
Part 2:

The last piece was bankroll management. I had to figure out what size the bets needed to be relative to my bankroll to be able to manage a long streak of losses. Using multiple runs of one million trials in my test program I found that a base unit of 1/200th of my bankroll reduces risk of ruin to about 0.25%. Rerunning my test prior to this post in 100,000 trials 281, or 0.281% failed. If success was winning 5 units and failure was losing my starting bankroll, in 100,000 trials I should have won 5 units 99,719 times (+498,595) and lost 200 units 281 times(-56,200) for a net win of 442,395.
The key to this system is cutting your losses so when you get to zero on a particular trial you quit and try again, relying on the 70+% win rate to carry you in the big picture.

As stated previously, I ran this for a long time with considerable success. It fell apart when I tried to apply it to basketball. I also realize now I shouldn't have been increasing my bets as my bankroll grew. I think I forgot part of my own strategy.

Were I to try this again, and I may since my sportsinteraction balance is still 9.96 and this was kind of fun, I would do it only on baseball, and only on games before the roster expands right before the playoffs. Pro sports may never recover from BLM so I might never get to try.

As I said at the very beginning, I know betting progressions shouldn't work. As a person with some science background, my thought was that if you have an idea that makes sense, you test it in a peer reviewed environment, even if prevailing logic says it is wrong. To those of you who chose to dump on me from the beginning and along the way, all I can say is I hope you find some kind of peace in your life to make you less miserable. For those of you who encouraged me, thank you. Whether it ultimately works or doesn't, it was a fun exercise that cost me nothing.
es330td
• Posts: 151
Joined: Mar 19, 2019
August 29th, 2020 at 1:23:32 PM permalink
For this part, I give credit to sentdex on youtube who posted the python code I modified to test this. You can view his video here: https://www.youtube.com/watch?v=bUFQNjjItok

#begin code
import random
import time

broke_count = 0

def reSort():
if len(system) > 6:
sum = 0

for x in system:
sum += x
sdiv = sum//len(system)
smod = sum%len(system)
for x in range(len(system)-1):
system[x] = sdiv

system[-1] = sdiv+smod
print(system)

def LaBouchere():
global broke_count
starting_funds = 200
goal = 5
global system
system = [1,1,1,1,1]

profit = 0

current_funds = starting_funds

not_broke = True

wins = 1
losses = 1

while profit < goal and not_broke:
if len(system) > 1:
size = (system[0]+system[-1])*2.5
else:
size = system[0]

if current_funds <= 0:
not_broke = False
broke_count += 1

elif current_funds - size <=0:
size = current_funds
not_broke = False
broke_count += 1

dice = random.randrange(1,101)

#assume 70% win rate
if dice < 30:
losses += 1
system.append(size)
reSort()
current_funds -= size
profit = current_funds - starting_funds

else:
wins += 1
current_funds += size
profit = current_funds - starting_funds

if profit != goal:
try:
del system[0]
del system[-1]
except:
pass

sample_size = 100000

for x in range(sample_size):
LaBouchere()

print("Broke Percentage:", ((float(broke_count)/sample_size))*100.0)
print("Broke Count:",broke_count)

#end code

If you aren't a programmer, sorry, I will not explain the code. I acknowledge errors of logic may exist in my code.
es330td
• Posts: 151
Joined: Mar 19, 2019
August 29th, 2020 at 1:27:21 PM permalink
Okay, that is everything. I look forward to hearing what people think of this.
ThatDonGuy
• Posts: 6414
Joined: Jun 22, 2011
August 29th, 2020 at 1:53:13 PM permalink
Quote: es330td

As I said at the very beginning, I know betting progressions shouldn't work.

Betting progressions work just fine if you have an Advantage Play.

If you can win a -250 bet 75% of the time, then, for every four bets, you win three, making \$300 profit, and lose the fourth, losing \$250, for an overall profit of \$50. Even if the losing bet is -300, you break even.
charliepatrick
• Posts: 2967
Joined: Jun 17, 2011
August 29th, 2020 at 2:02:47 PM permalink
If you've found a method of identifying odds-on bets of -300 (1/3) or better having a positive expectation, then just make your wagers on them and, in the long term, you should come out ahead. Some people here will be able to advise on how to manage your bankroll better than I can.

Conversely any system on -EV bets cannot win in the long run. With betting progression they may win gradually for a long time but eventually the long-shot of consecutive losses wipes them all the profits out.
redietz
• Posts: 767
Joined: Jun 5, 2019
August 29th, 2020 at 2:33:59 PM permalink
Let me fire from the hip here quickly.

If you're not a savant regarding a particular sport, the progressions don't mean anything. I'll get back to that in a minute.

Now, this is a complicated subject, because if you are a savant and you start off a season knowing the teams, then as you have more success, the implication usually is that since you have demonstrated that you know the teams, then amping up wagers after you've had success may make sense. However, if you are using certain angles or if you wait for certain perceptions of values accrued before betting, then after you've had early success, those angles or values accrued may be exhausted, so even though you know the teams better as the season progresses, the values are gone, so progressions are a bad idea.

Starting from the basic idea that you are not a savant for a particular sport, which covers all but about a hundred people in this country, a progression isn't going to have any effect.

Basic rule of thumb -- if you actually know what you are doing, there is no need for progressions or parlays or gimmicks. In a sense, those are distractions that people sometimes use as excuses. The gimmicks muddy the waters of whether someone can win or not. Now I vary bet sizes, quite a bit actually, although considerably less than 30 years ago. But lifetime, my unit percentage is less than one percent better than my game percentage, so my varying of bets, if I were criticizing myself, may be a waste of time best spent on other kinds of analysis. I perceive a big-ass difference between the quality of certain kinds of plays, but long-term, that perceived "monster difference" is really just one percent.

I have a few more comments, but have some work. Will check in later.

Bottom line: even though sports betting is a "game of opinion," and therefore can be beaten, progressions really have no place in trying to win long-term. Historically, those people who have run roughshod over certain sports for periods of time, I really don't know anyone who relied on formal progressions of any kind. People who were killing it and knew what they were doing (say, Billy Walters and the Computer Group around 1990) increased their wagers as their bankrolls grew. The way I do things, I increase wagers as I win and decrease them when I lose, because I use a percent of bankroll rating for games. But formal progressions? No -- of no value.
"You can't breathe dead hippo waking, sleeping, and eating, and at the same time keep your precarious grip on existence."