Board Thread:General Discussion/@comment-2158701-20140130092413

Just a heads up, the following post will be long and very technical in nature, and will be a summary of the tons of errors and inaccuracies on the IC2 pages, especially regarding machines and overclockers. Through this post, I hope to increase understanding of how EU and overclocker game mechanics really work, so it can be possible to organize a group effort to find, test, and correct all such errors.

One error that is near omnipresent in all IC2 machine pages is how it lists the max eu input. The max input a machine can take is NOT eu/t, which is current - it is actually eu/p, (eu per packet), which is voltage. Current and voltage are NOT the same thing, and about every page lists eut instead of eup. For instance, it is easily possible to output 250,000 eut, in 32 eup, if you transform all the power down to LV first, and that will never blow up a machine as it is within the max acceptable voltage of 32 eup.

The values for overclocked machines are a particularly nasty problem, which thanks to how short a game tick is, is time consuming to get the processing time values to the exact tick. The method would involve processing a large amount (several stacks possibly) of items, then dividing by time. Trying to count or estimate the value PER operation is totally inaccurate. There are a ton of inaccuracies with overclocker values, and it doesn't seem like the wikia has any accurate information to how the game mechanics actually work on them.

For example, the overclocker values for Electric furnaces are DEFINITELY wrong. Please scroll down to the values there before you read the following analysis. According to the theory there, "...the values are rounded DOWN to the nearest tick and EU/t respectively)"

Assuming 8.75 s is indeed the processing time with no overclockers, 8.75 * 0.7^10 = ~0.247, which according to the theory, would round DOWN to .20, not .25, and I have confirmed in game that the processing time is indeed .2, or 5 per second, not 4.

Now, if it weren't for another error, I would've concluded that the values were just sloppily rounded to the nearest hundredths digit instead of rounded down to the nearest tick, but the other error is, the actual processing time of an electric furnace is much closer to 6.5 seconds than 8.75 seconds, which throws that entire table off. with the only saving grace that values are rounded to a tick (.05) instead of hundredths digit (.01), causing slightly incorrect non OS values to still converge to a single value when multiplied by exponentials like 0.7^10.

The overclocker value for 10 overclockers assuming 6.5 seconds, is now ~0.1836, which if you round up or to nearest, would give the same value as .247 rounded down!

A set of values that does not agree with the "round down to nearest tick" theory are the OC values in macerators, extractors, and compressors. A timed test agrees with the default 20s per operation with no OC (phew), but now, let's look at 15 overclockers. That would return a value of roughly .095, which would round down to .05, and so would 16 overclockers at 0.0665. Of course though, 15 overclockers is actually twice as slow as 16 overclockers, processing at .10 (2 ticks) per operation, compared to 16's .05 (1 tick) per operation.

My theory for this then, is that the calculated value actually rounds to the NEAREST tick, not rounds down. That is the only explanation I see for .095 rounding up to .10, with .0665 rounding down to .05.

Every page on IC2 machines needs to get those values corrected, as the only values posted are either completely wrong (8.75s for electric furnace, what...) or erroneously rounded to the nearest hundredths digit instead of nearest tick. This will of course take actual testing to verify that the math agrees with the results, and the larger sample data, the better, thus requiring more people to test this.

Those are the two main categories of errors I've found. There are also just a ton of sloppy errors caused by a lack of testing, such as scrap with mass fabricators. They don't actually reduce energy usage by 1/6 and increase speed by 6x, but rather only UP to those values. Powered by 65536 eut, the speed increase is reduced to only ~50%, and at 1 million eut, 0%, or no increase at all. And this is of course due to the operations/tick concept. Increasing speed has to reach the breakpoint of the next tick, not just the next hundredths digit.

I hope editors will join me to really reverify and test all the IC2 pages, because what is posted and presumably accepted as the correct values are way off. I've caught several, but I can only read and correct so many pages. The brewing page was one such page that I corrected that seemingly just had completely made up values for the ratios, and if I can find so many errors in such a short period of time, I believe most the pages here really need to be reverified for accuracy. 