Donkey Kong Questions

[kalel]

I apologize if any of these questions have already been discussed.( i couldnt find them )  #1. What is the theoretical max and legit max . Is anyone else besides lakeman and mccrurdy. Trying legitly to practice and get good enough to get there?( i would like to watch their streams ) and Has anyone done the extensive math on those maxes ?

[xelnia]

I apologize if any of these questions have already been discussed.( i couldnt find them )  #1. What is the theoretical max and legit max . Is anyone else besides lakeman and mccrurdy. Trying legitly to practice and get good enough to get there?( i would like to watch their streams ) and Has anyone done the extensive math on those maxes ?

If you take the absolute best scores ever achieved on every board type and apply that to a whole game you get something like 1,812,300 as a theoretical max.

If you take the best Start, best full-game board averages, and best deaths you get something like 1,309,900.

If you take the best Start, best level average, and best deaths you get 1,285,800 (McCurdy has all these records).

If Robbie could replicate his best Start, best full-game board averages, and best deaths in one game he would score 1,275,900.

If Robbie could replicate his best Start, best full-game level average, and best deaths in one game he would score 1,260,700.

If John could replicate his best Start, best full-game board averages, and best deaths in one game he would score 1,305,800.

Personally, I think the community should shift away from the "what is the max score?" question and look more towards the odds of getting a particular score. There's way more data for John than any other player and if we assume that he can get a 140k Start and 30k in deaths in any given game (not a great assumption, but let's work with it), he would still need a 65,000 level average for 1.28M...and that's probably a 1 in 300 killscreen game for him. That's a rough calculation but it makes me think that a 1.28M is still a ways off.

Of course this all takes into account how people are currently playing. If some new strat is found/implemented then everything could change.

Other players currently playing at a high level are Thomas Bauer (https://www.twitch.tv/kig666) and J.P. Buergers (https://www.twitch.tv/wflimusic). J.P. just had a 1.132M game that would have been 1.15M if he had even decent rivet luck. Thomas' best is 1.131M and grinds daily looking for a 1.15M score. I'm not sure what Thomas' ultimate goal is, but I think J.P. is looking towards the top of the leaderboard. Obviously Dean could strike with a big score any day. And I think Wes is getting back into DK looking for a PB.

[kalel]

 Thsnk you . Exactly what i was looking for 😀

[Barra]

Thsnk you . Exactly what i was looking for 😀

1.5m junior go

[homerwannabee]

By the way, with McCurdy it's obvious the odds of a 1.3 million game is better.
Are the odds better than a billion to one at least?

[jwade614]

I created 10,000 simulated killscreen games based off John's games on kongtracker going back to when he first got the record.

I looked at full, completed levels (not individual boards) and deaths. Data consisted of 9 starts and 109 levels 5+.

Based on this, odds of 1.3 is about 1 in 1000. I plan to revisit this as more data comes in.

Score   Games   Odds
1.185   1   100.0%
1.190   0   100.0%
1.195   4   100.0%
1.200   9   100.0%
1.205   23   99.9%
1.210   57   99.6%
1.215   110   99.1%
1.220   202   98.0%
1.225   336   95.9%
1.230   572   92.6%
1.235   774   86.9%
1.240   1068   79.1%
1.245   1233   68.4%
1.250   1323   56.1%
1.255   1218   42.9%
1.260   1049   30.7%
1.265   799   20.2%
1.270   522   12.2%
1.275   351   7.0%
1.280   192   3.5%
1.285   89   1.6%
1.290   44   0.7%
1.295   15   0.2%
1.300   5   0.1%
1.305   4   0.0%

[homerwannabee]

 Wow, only 1 in a 1,000 killscreen games at his high level of play. A lot closer than I thought.

[Barra]

Wow, only 1 in a 1,000 killscreen games at his high level of play. A lot closer than I thought.

That's still a killscreen game a day for almost 3 years.

[homerwannabee]

Yep, but the odds are now not mind numbingly insane.  If he were to get two killscreens a month for five years, that's a 12 percent shot.

[xelnia]

I created 10,000 simulated killscreen games based off John's games on kongtracker going back to when he first got the record.

I looked at full, completed levels (not individual boards) and deaths. Data consisted of 9 starts and 109 levels 5+.

Based on this, odds of 1.3 is about 1 in 1000. I plan to revisit this as more data comes in.

Score   Games   Odds
1.185   1   100.0%
1.190   0   100.0%
1.195   4   100.0%
1.200   9   100.0%
1.205   23   99.9%
1.210   57   99.6%
1.215   110   99.1%
1.220   202   98.0%
1.225   336   95.9%
1.230   572   92.6%
1.235   774   86.9%
1.240   1068   79.1%
1.245   1233   68.4%
1.250   1323   56.1%
1.255   1218   42.9%
1.260   1049   30.7%
1.265   799   20.2%
1.270   522   12.2%
1.275   351   7.0%
1.280   192   3.5%
1.285   89   1.6%
1.290   44   0.7%
1.295   15   0.2%
1.300   5   0.1%
1.305   4   0.0%

Are you using R for this? If so, would you mind sharing your code?

[TheSunshineFund]

A python is a 

[Weehawk]

<Sonium> someone speak python here?
<lucky> HHHHHSSSSSHSSS
<lucky> SSSSS
<Sonium> the programming language

http://bash.org/?400459

[jwade614]

Are you using R for this? If so, would you mind sharing your code?

No. I did it in an OpenOffice Calc spreadsheet using the rand() function. Lots of copy/paste and manipulating data. When I have the time and I feel like it, I'll figure out how to write macros in OpenOffice Calc, and I'll simulate based on boards instead of levels.