Balloo85 5 Posted February 14, 2020 Earlier today I did a small sample of 200 rolls, put it all in a spreadsheet and did a chi-square test on it. Just to see how balanced it really is. Out of this (very small sample) it really does seem pretty balanced. https://www.scribd.com/document/446968608/Trust-Dice-Analyze 2 Quote Share this post Link to post Share on other sites
Hippos4ever 6 Posted February 14, 2020 Yeah, looks about right. Quote Share this post Link to post Share on other sites
Balloo85 5 Posted February 14, 2020 The problem is when you bet for example on 66-100, that you can have a bad streak on both the middle numbers and the low numbers in a row. That's what's really screwing you over sometimes. Quote Share this post Link to post Share on other sites
Balloo85 5 Posted February 14, 2020 The second test doesn't look good at all. It's also from 200 rolls, but this time I was betting BTC instead of TXT. Are the dices programmed to be nicer when betting TXT? Results: https://www.scribd.com/document/447011214/Trust-Dice-Analyze-BTC Quote Share this post Link to post Share on other sites
Balloo85 5 Posted February 14, 2020 12 hours ago, Balloo85 said: The second test doesn't look good at all. It's also from 200 rolls, but this time I was betting BTC instead of TXT. Are the dices programmed to be nicer when betting TXT? Results: https://www.scribd.com/document/447011214/Trust-Dice-Analyze-BTC The chi-square test in this one is wrong. It assumes that there are equal chance to hit 0-35 as 36-65 and 66-100. The chances are of course 35-30-35. Here's how this test shows in reality. The hypothesis that the distribution of 0-35 rolls and 65-100 roll are rejected. With a statistically significance of 0,05 1 Quote Share this post Link to post Share on other sites
Balloo85 5 Posted February 14, 2020 (edited) Expected outcome would be 0-35 (35%) 36-65 (30%) and 66-100 (35%) Real outcome: 0-35 (30,5%) 36-65 (35,5%) and 66-100 (34%) It is a relatively small sample, given how many bets I make per day. But it should be big enough sample to at least give a hint on how it look overall. Most notable is how the 36-65 rolls happens 5,5% more often than expected. Which is also the numbers people are less likely to bet on. I should state that this thread isn't meant to complain and moan. But the more we know, the wiser we can place our bets. Edited February 14, 2020 by Balloo85 Quote Share this post Link to post Share on other sites
Balloo85 5 Posted February 14, 2020 It's been a while since I had to work with statistics, so this one might be wrong. If anyone see anything odd with it, let me know. Quote Share this post Link to post Share on other sites
Balloo85 5 Posted February 14, 2020 Although this makes me wonder wtf is going on. No one in their right mind would assume something like this happens by chance. Quote Share this post Link to post Share on other sites
Laura 20 Posted February 15, 2020 7 hours ago, Balloo85 said: The chi-square test in this one is wrong. It assumes that there are equal chance to hit 0-35 as 36-65 and 66-100. The chances are of course 35-30-35. Here's how this test shows in reality. The hypothesis that the distribution of 0-35 rolls and 65-100 roll are rejected. With a statistically significance of 0,05 Hey Man, you're genius! Unfortunately, my math is bad, so I may not clearly understand all your points. What I know in Casino industry is that all random results follow the "law of large numbers" and you need a rather large sample to make the law work. Besides, all the licensed casino will not manipulate result, as they value reputation more than short-term return. For example, we paid 100k+ EUR in order to get the license. Why we do that? we believe that a regulated casino will have better reputation and attract more players in the long term, and thus bringing more value to the TXT eco-system. Quote Share this post Link to post Share on other sites