User talk:Captain deMarco

From Star Trek Timelines Wiki
Jump to navigation Jump to search

Welcome


Hey Captain deMarco, welcome to the wiki. Thank you for your contributions to our crew page. -- ST Logo.png MilkN2Sugars (My talk page) 12:16, 17 January 2016 (MST)

Template:DropTest/Death in Battle/1


Just letting you know that I switched you to first credit in Template:DropTest/Death in Battle/1 because you have more runs. Also note that your position in the #expr tags is also now in front as per Template:DropTest#Tips and guidelines. Thanks for you contribution! — CodeHydro 13:26, 14 June 2016 (CDT)

Template:DropTest/Hot Pursuit/1


Please don't remove old data without evidence that it is actually invalid. Compare your June-only version to the version with all the data. Notice how June-only has a lot more red numbers than the full version. Red means "statistically useless". Hover your mouse over each number and you can see the margin of error for the costs. Notice how the full data version's numbers fall completely within the margin of error of the June-only data version.— CodeHydro 08:51, 15 June 2016 (CDT)

I just analyzed the May-only and the June-only data sets and found an average of 79.9% overlap in the range the three "fairly reliable" items. That means there is a nearly 4 times higher chance that the two data sets are one an the same versus the chance that they are actually different. — CodeHydro 09:20, 15 June 2016 (CDT)

If you say so. I am not very experienced as a wiki-contributor and don't know what the do's and don'ts are. If I did a don't, I apologize. The idea behind my action was that as of june 1st, droprates have been adjusted by DB and therefor older results are unreliable. I expect this will happen more often. Still think it wouldn't hurt to get rid of old data every once in a while. But I'll restrain myself. Excellent work with creating these templates by the way. - Captain deMarco

  • Thank you. Yes there are procedures for removing data. Eventually we may wean off of pre-June data, but right now it's too soon and so far out of the dozens of missions reviewed we haven't find a single conclusive case where pre-June actually has a different drop frequency despite the reduced quantities of components. No worries however. Don't worry about the inexperience. We were all there at one point ;) Even I get nudged by other experienced editors now and then. — CodeHydro 09:25, 16 June 2016 (CDT)

Template:DropTest/A Tale of Forgotten Lore/1


FYI, I moved you to by2 in Template:DropTest/A Tale of Forgotten Lore/1. Also note that we do not generally split up samples from the same tester, i.e. if you did warp 10 twice, just add them up (raw data will show non-added version if you want to share). Basically we want the by-parameters to match up with the relative position in the #expr tags as it makes it easier for other testers to figure out where their old data was. In the above case, you may have seen 3 items in the expr tag but only 2 tester; that's because one tester forgot to give himself credit. Thanks for sharing! — CodeHydro 08:57, 16 June 2016 (CDT)

  • PS: if you really want to split up your samples for some reason, just put all your results in a parenthesis so that everyone knows that group of samples is yours. — CodeHydro 08:59, 16 June 2016 (CDT)
    • Excellent and elegant. Will do. - Captain deMarco 0:36, 17 June 2016 (GMT+1)

Audit on Template:DropTest/Mutual Assured Destruction/1


Regarding Template:DropTest/Mutual Assured Destruction/1, I noticed that you added 7 runs with these two edits but eventually only subtracted 2 runs when segregating your data sets here. In the current version of the page, I inserted the 5 undocumented runs in the front of your parentheses. Please double check your edits in the page history and your own records to make sure I interpreted your edits correctly. — CodeHydro (msg) 21:13, 4 July 2016 (CDT)

  • Could well be. This was one of the first edits I did regarded drop rates. I don't recall exactly what happened, nor do I keep personal records. I think your assumption is probably correct, but for the sake of accuracy you might want to consider deleting the 5 undocumented runs. — Captain deMarco (msg) 17:09, 5 July 2016 (GMT+1)