NOTES · DIVIDED VOTES
Most group processes dey treat divide as result. Half of di room tok yes, half tok no, di motion fail, and di question disappear from di record like say dem don answer am. For causal question, dat one dey mislead: di room no decide say A no get effect on B. E no decide anytin at all.
True Causes dey treat divide anoda way. Wen panel deadlock on whether one cause get big hand for anoda one, di question no dey fail quietly. E dey stay open and e go anoda place make dem settle am. If second panel divide too, dem go publish di pair as argument and e go stay visible for di map. Di disagreement na part of di result.
To ask di same pipo again no dey usually change di answer. So question wey divide dey comot from di group wey divide on am and e dey go to two oda whole groups, wey dem choose so nobodi fit steer am go meet friendly audience. Dem dey answer separately, and dem must gree. Di way dem dey share questions dey rotate over time, so no group go own one area of di map.
Dem add di two rules becos one simulated bloc break di version wey no get dem. With escalation, di method dey make about 3 wrong links for evri ten thousand questions wey dem decide, even wen each pesin dey answer 30% of questions wrong. With referral, room wey get three seats wey dey vote togeda dey move from recovering 0.43 of di real drivers to 0.98.
Di principle dey general. Any process wey dey turn “we no fit gree” into “no” dey let small group remove cause just by dividing di vote. To keep di argument for di record, and send am to pipo wey no dey di room, dey prevent dat one.