Hikaru Nakamura vs Vladimir Kramnik ANALYSIS

I decided to investigate the controversy myself. The data I collected is from a SIMULATION. It is NOT from Hikaru's actual account. I remained objective in my analysis and attempted to show both sides. Please spread the word on this analysis. It would be amazing if it were shared with others including those directly involved in the controversy. I hope my analysis helps everyone get to the right answer. At no point in the analysis is judgement assigned to either party. The whole intention was to analyze what could happen in a simulation. 

In my simulation, 2,000,000,000 total games were played. Hikaru's ELO was between 3150 and 3350, and his opponents ELO's were between 2700 and 3100. The results were fascinating. Please check it out by clicking the link to my GitHub Project below. 

Feel free to give me any feedback on my work. To be clear, I did this for FUN, not for profit. I just love chess, and analyzing simulations.

Vladimir vs. Hikaru

Love me some python. Code looks good. Outcome is as expected not good :(

Hey, this seems fascinating. But can you please update your Conclusion section - currently it does not summarize the outcomes, or draw any conclusions as a result.

Many users (such as myself) will look straight to the Conclusion section first, to see what the result of the analysis has been.

Good job on all the work you put into this.

Interesting. If I'm reading it correctly, then basically a 40 win streak wouldn't be that unusual for Hikaru against these lower rated players.

I have just updated my conclusion as requested. If you disagree, please share your perspective, and let's have a conversation about it. This was very fun for me.

We all live in a simulation, none of this is real!

Hey, that is interesting! I've been doing very similar calculations over the last couple of evenings and just published this on my blog last night:

Analyzing abnormally long win streaks in online chess

The methodology is very similar, although I have taken the probability slightly different. This isn't clean, but I think I'll refine my analysis if I have more time this week and calculate with more precision.

Simulation is a serious tool IMHO. I would be more inclined to tackle this kind of problem using a script. Partly because statistical methods can be rather complex and some of the maths is a bit beyond me right now. However billions of simulated game results ought to reveal typical trends. The algorithms required to generate realistic outcomes need not be complicated (a good pseudo-random number generator and a few correctly weighted arguments ought to do the trick). I was just looking at the Read Me file. I'll enjoy reading this and I'm sure I'll learn some new things in the process. Looks excellent BTW - Thanks!

Bro jumped right into conclusions

https://www.chess.com/forum/view/general/sign-my-petition-to-defend-our-great-hero-kramnik

UPDATE:

I am going to do the same simulation, but MUCH bigger. I will be doing this for ALL players that can be rated from 100 all the way to 3700. When that data is complete, I will upload two different datasets that can be analyzed by anyone. The first will be the whole dataset top to bottom, and the other will be the whole dataset excluding matches where the difference in ELO between the two players is greater than 500. Let me know your thoughts on this idea.

@michaelpowers8 You could also include losing streaks. I've had some of those. Nowadays, I only play a few games per session. This avoids losing too many points when I'm tired. I generally don't find time to play much, but it does help with stress relief when I need a break from the work I do.