As for the "format", this was actually borne out early in in reviewing DK and developed over the various submissions by Steve.
A couple of key statistics were staples of all WR articles that we typically provided to gamers...1st life total, how fast to 1M, each game death, etc. But for DK I thought it was time to do something special. The game was finite, like Pacman, but there was no "perfect score". So the thought occurred to me as you may see at the end of my articles, what was the "theoretical maximum" for a game of DK ? And here's the though process behind the initial format which has since been fully blown out by gamers here on DKF.
1st - there was a kind of "myth/legend" about the 1st-life score after stage 2-3. Bill said that years back he had one game where he finished that stage with 50,400 points which, at the time, was an unheard of threshold. So that was my first important benchmark when determining the score. I took what Steve did and deducted it from 50,400 and thought that whatever his final score was, that difference in theory could be achieved on top of that
2nd - same logic was applied to end of stage 4-5, the last time before the format was all 6-stages per level and 8000 bonus per stage. Using empirical data from the era based on the few attempts chronicled in depth, this was the next benchmark...since Bill had mentioned how many of his own games were aborted at that point if he was not on pace for (I forget) either 130K or 140K. Adjusting for the 2-3 differential between 50,400 and the score that game was subtracted from the differential thru 4-5.
3rd - taking a simplistic approach, it seemed reasonable to add on 51 barrel stages, and 17 each of the other three board types, based on the highest validated scores from all of the combined and validated submissions up to that point. Sure, latter stages were harder, but at that time this was the inception of the "theoretical maximum" calculation so a more refined methodology had not yet been envisioned
4th - next, the concept of the extra lives, and the "what if" element..."What if someone could reach stage 115 on their 1st life ?". I thought at the time based on what I had seen that the optimal strategy would be get to 115, the final barrel stage, point-press each to the maximum possible on each life (at the time 9500 points per life) then finish up the final rivet stage and reach the kill
5th - and finally, the score possible with stage 22-1. It was initially debated between us what was possible in this stage...I forget why but 1100 points and 700 points were the initial marks used as it was not clear just how much someone could get with the maximum amount of luck.
6th - adding up the above factors, that was the original calculation for "theoretical maximum"
Afterwards, each time a new "higher benchmark" was achieved for any of these thresholds...2-3, 4-5, individual stage types from levels 6-21, or the kill, the "theoretical maximum" would be adjusted. Thus every time someone pushed a barrel stage by 100 points that means +5100 theoretical maximum points...and +1700 for any of the other stage types if they were surpassed.
The very first calculations I discussed privately with both Bill and Steve, and I might have even made mention of them within one of the articles. A score of 1.30xM was determined to be wildly possible but not likely...statistically no one could pull off EVERY optimal element at once...capturing all the fireballs in every rivet stage, leeching in every elevator stage, no time wasted in a conveyor stage, etc. So a more plausible 1.200M was deemed as "do-able" with some luck, while up to 1.250M was do-able with "a LOT of luck".
Ironically one of today's calculations that I read suggested 1.265M was do-able, and Robbie told me that included the extra approx. 10K that MAME DK allows, so 1.255M for arcade...very close to my 10+ year earlier estimation !! At the time we had no idea that MAME DK offered a different point-based opportunity that arcade DK did.
Anyway, that's the story of how the original format of the "theoretical maximum" score for DK was developed and enhanced between 2002 and 2004.
On an interesting sidebar...
After the two 1M scores were submitted by both Bill and Steve, I told them each between 2004 and 2005 what I thought was possible from each of them.
-> To Steve I estimated that 1.050-1.080M was possible if he learned new tactics. Steve was in my opinion a hammer-player while Bill was a jumper, and it seemed that jumping was far more lucrative that simply hammering
-> To Bill I estimated around 1.123-1.127M based on two earlier performance...the 1,047,200 that Bill previewed at Funspot in May/05, and based on what I heard he had done from an earlier live performance.
For those who do not know, Bill's 1,047,200 performance was intentionally 100,000 even above Steve's much earlier 1st TG recognized WR of 947,200 done via DDK, so you knew coming into it this was intentional, or rather once you saw the final score. Anyway, Bill started dumping points like crazy when he neared the Level 20 elevator stage and he let the timer run out substantially before climbing the ladder and making the score "all zeroes"...the infamous glitch that was shown at Funspot since Bill sent the 2nd gen copy instead of the original.
Bill then went from that point to the kill and only scored 47,200 more points which is ridiculously low AND he killed off I think 2 men in the process at the end. When calculating based on his other validated benchmarks this performance had the chance to be in the 1.123-1.127M range.
On a footnote, this was the score that Walter stupidly entered on camera for KoK even though when we first watched it in the cabin we agreed it could not be accepted. Those of you who have the KoK DVD with the director comments can listen to Seth and Ed talking about how I personally reversed that score entry after coming back from Funspot...info that the average movie goer never knew but which only those who bought the DVD AND listened to this would find out. But because KoK was a crafted tale, this element was intentionally left out of the final product. Judge for yourself.
Bill had another performance which I did not see but only heard of in which he intentionally walked away from the machine way early in the game, and based again on previous benchmarks achieved by Bill, that one also had the potential for 1.123-1.127M. So Bill was told by me that was his likely achievable range...roughly 50K more that Steve, unless Steve learned some new tactics.
After that came Steve's 1.049M performance which is another story !!
