Update index.html

new schizo of the week .. please dont crash the site ...

Author: tumultuous0

index.html

@@ -675,20 +675,15 @@ <h2 class="text-center"><img src="gifs/1570998246941.gif" height="75px" alt="">L
     <hr>
     <br>
     <div class="rant">
-      <h3 class="text-center"><a href="others/schizo.html">Steve of the Week</a></h3>
+      <h3 class="text-center" style="font-family: 'Times New Roman', Times, serif; color: #fc8a8a;"><a href="others/schizo.html">Hopeless Mortalbot Schizoparanoia of the Week</a></h3>
       <center>
-        <p>Hey, fellow Stevefags, how's it going? Just another few hours until we get ballerina Steve announcement. Oh, what's this? Some new tweet? Sounds like it'd be good for confirma-</p>
-        <p>"[New Outfit - Pointe Monologue - Mirai]"</p>
-        <p>AAAAAAAAAAAAHHHHHHHHHH</p>
-        <p>OH NONONONONONONONONONONONONONO NONONOOOOOOOOOOOO!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!</p>
-        <p>THE T....TWEET IS FAKE THEY DREW MIRAI THEY........THEY CAN'T DO THIS BROS FUCK MIRA FUCK HAGS FUCK YOUR STUPID BOOMER IRRELEVANT BULLSHIT ITS NOT FAIR</p>
-        <p>CATFOOD ITS NOT FUCKING FAIR WHERE IS MY STEVE WHERE IS CALADBOLG WHERE IS THE CRIMSON CHEMIST OH NO NO NOOOO OOOO</p>
-        <p>PLEASE GOD ANYTHING BUT THIS PLEASE DON'T LET DADDY YOSTAR BE WRONG PLEASE NO PLEASE OH MY GOD NO NON ONONONONOO</p>
-        <p>Y...YOU GUYS LOOK AT MIRAI'S DRESS HER IT'S FUCKING TRANSPARENT WTF IS SHE A LITERAL STRIPPER????? LITERAL WHORE OUTFIT???? NOOOOOOOOOOOOOOOOOOOOO</p>
-        <p>ITS FAKE ITS GOT TO BE FAKE ITS REALLY REALLY GOT TO BE FAKE GUYS LISTEN PLEASE LISTEN BALLERINA STEVE IS IN THE GAME!!!!!!!!!!!!!!!!!!!!!!!! BALLERINA STEVE IS IN I FUCKING PROMISE BALLERINA STEVE IS IN YOU ARENT GETTING MIRAI YOU ARENT GETTING YUZU YOU ARENT GETTING CHUUNI OR BUNNY OR</p>
-        <p>CATFOOD WHAT THE FUCK I TRUSTED YOU WHY DID YOU FUCK IT UP WHY DID YOU MAKE MIRAI'S OUTFIT GOOD NO THIS DOESN'T FIT MY NARRATIVE THAT YOSTAR IS RETARDED NO THEY CANT BE COMPETANT THEY JUST CANT</p>
-        <p>FUCK YOU FUCKING TWINTAILS WHORE FUCK YOU FUCK YOUR STUPID DRILLS I HATE YOU I HATE YOU MIRAI FUCK YOU YOSTAR FUCK YOU CATFOOD FUCK YOU!!!!!!!!!!!!!!!!!!!!!!! FUCK YOU FUCK YOU FUCK YOU NONO NONONONONO</p>
-        <img src="https://files.riichi.moe/mjg/game%20resources%20and%20tools/Mahjong%20Soul/game%20files/emotes/kristeve-8.png" height="100" alt="stevebros..." /> 
+        <p>I had suspicions early on. M1 to Saint in 190 hanchan? -- come on, kid. He tried to cover his tracks, but rumors circulated. Jane got him in the end.</p>
+        <p>I was tipped off by something else. During the ascent, his win rate was 22.6% (74th percentile) and his deal-in rate was 8.7% (1st percentile). The delta of 13.9% is impressive - but to be in the bottom 1% of deal-in rate and the top 26% of win-rate simultaneously is inhuman, since, among good players, <a href="https://pathofhouou.blogspot.com/2021/02/analysis-houou-stat-correlations.html" target="_blank" rel="noopener noreferrer">the two stats have a strong positive correlation</a>. In other words, low deal-in rate players are more likely to have low win rates, and high deal-in rate players are more likely to have higher win rates (well, assuming they're not a looper).</p>
+        <p>Seeing the cat get his comeuppance validated my stat-based suspicions. It also made me wonder if mortalbotters could be distinguished via literally just these two stats. Is there a cluster of mortalbotters roaming Jade with extremely low deal-in rates and unusually high win rates (at least, relative to the rest of the Jade Room players)?</p>
+        <p>Well, let's see. I scraped a few thousand profiles for Jade Hanchan players with >100 games and ended up with >1000 entries for each rank of M2-S3. Mortalbots should show up on an X-Y scatterplot of win rate vs. deal-in rate as a small smear in the bottom right hand corner (high win rate, low deal-in rate).</p>
+        <p>Hey, <a href="https://files.riichi.moe/mjg/oc/images/General/Plot-Autism/ms_facet.png" target="_blank" rel="noopener noreferrer">look at that</a>. It's not obvious in the M2/3s, but among S1/2/3s, there's a subpopulation centered around (24, 10) distinct from the larger cohort centered around (22, 12). A classifier places ~5% of the Saint players <a href="https://files.riichi.moe/mjg/oc/images/General/Plot-Autism/classification.png" target="_blank" rel="noopener noreferrer">into this subpopulation (represented in blue)</a>.  Unsurprisingly, these extraordinarily high-win:low-deal-in players tend to have <a href="https://files.riichi.moe/mjg/oc/images/General/Plot-Autism/wde.png" target="_blank" rel="noopener noreferrer">high net win efficiencies</a> (win rate*average win score - deal-in rate*average deal in score) and <a href="https://files.riichi.moe/mjg/oc/images/General/Plot-Autism/14_rate.png" target="_blank" rel="noopener noreferrer">great placements</a> (high 1st rate, low 4th rate).</p>
+        <p>Are they bots? In a luck-based game, you can't be certain. Glancing through Amae profiles from a dozen or so suggests most are (based on queue behavior). Maybe some additional investigation is in order.</p>
+        <img src="https://files.riichi.moe/mjg/game%20resources%20and%20tools/Mahjong%20Soul/game%20files/emotes/ninomiyahana-2.png" height="100" alt="detective love" />
       </center>
     </div>