Talk:Repairs/@comment-4148196-20110720002435

Method: I built 15 brand new, level 0 hot dog stands. I never upgraded them. I then collected them directly on the green buck (Figured I might as well knock off several hundred against that trophy simultaneously). When one broke, I fixed it and repaired it quickly with cash (not piastres) so I could collect all 15 each round.

Assumptions: 1. The level of the building does not effect the repair rate. - May be wrong, is easily testable, but for now I went with what's simple. 2. When you collect (yellow buck, green buck, or longer) does not have an effect. - This is also made for simplicity's sake. If it's wrong the testing gets MUCH more complicated.

My theory is that there is an increasing probability that a facility breaks each time you collect from it. When the facility breaks, the probability function gets reset. My data show a strong liklihood that something like this is happening. Here's the raw data: x = #, #, #, #, ... etc. x = collection number after being built
 * 1) = 1 through 15, I tracked which of the fifteen hot dog stands broke at each collection.  A zero means none broke that collection.

1 = 0 2 = 0 3 = 0 4 = 0 5 = 0 6 = 0 7 = 0 8 = 0 9 = 0 10 = 3  -This was a surprise that it took this many iterations for one to break. 11 = 12 12 = 5, 6 13 = 13 14 = 1, 8, 10, 11, 15 -here I got excited. It seemed like the probability of them breaking was definitely increased, though I do realize that this isn't truly enough data yet. 15 = 2, 4, 7, 14 -only one left, #9, I felt it would be seriously interesting if it broke next and it did 16 = 9 17 = 0 18 = 0 19 = 0 20 = 0 21 = 0 22 = 11 -Interesting. Only eight collections had past since 11 was repaired. 23 = 8, 13, 15 24 = 3, 6 25 = 5 26 = 10 27 = 7, 12 28 = 0 29 = 1, 4 30 = 0 - At this point, only 2, 9, and 14 hadn't broken twice, and it's been eight collections for #11 again so it should be available to be broken. 31 = 14 32 = 0 33 = 2, 9 -note how the last ones to break a second time were towards the end of the first round of breakages Here I was getting bored (This was an hour's worth so far). And since there had not been any overlap yet (i.e. no third breakages) I thought this was a good place to let it wind down. I stopped doing repairs. If a building broke, I demolished it, so after they start breaking again, each collection will be less than 15 total collections. 34 = 0 35 = 15 36 = 6 37 = 8, 11, 12 38 = 10 39 = 4, 7 40 = 3, 5 41 = 1, 3 42 = 0 43 = 2, 14 44 = 0 45 = 0 46 = 0 47 = 0 48 = 0 49 = 0 50 = 9 And that's all of them broken three times each and demolished. Total collections = 596 Repairs needed = 45 45/596 = 7.55% (My lifetime collection/repair rate is 4055/58245 = 6.96%)

Here's how many collections each individual stand took to break: 1 = 14, 15, 12 2 = 15, 18, 10 3 = 10, 14, 16 4 = 15, 14, 10 5 = 12, 13, 15 6 = 12, 12, 12 7 = 15, 12, 12 8 = 14, 9, 14 9 = 16, 17, 17 10 = 14, 12, 12 11 = 14, 8, 15 12 = 11, 16, 10 13 = 13, 10, 18 14 = 15, 16, 12 15 = 14, 9, 12

Longest stretch without breaking = 18 collections Shortest stretch without breaking = 8 collections

Collections until breaking after repair or being built: 1-7 collections = 0 8 collections = 1 9 collections = 2 10 collections = 5 11 collections = 1 12 collections = 11 13 collections = 2 14 collections = 8 15 collections = 7 16 collections = 4 17 collections = 2 18 collections = 2 19 and more collections = 0 This does assume that being built and being repaired reset the probability the same way. Note, I believe this would probably make a pretty bell curve with enough data points. I suspect the low incidence of 11 and 13 collections to be a fluke due to there only being 45 data points.

So I think this shows that there is an increasing probability of breaking which averages over the long term around the 14th collection. The next question is, is there a 0% chance of repairs needed for a number of collections (like 5-7), or is the chance of it breaking just too small to have showed up in this limited run? It also certainly seems like there is a strong enough increase in chance of breaking after the 14th that it would be a rare building to go 20 collections without breaking. Anybody got enough statistical expertise to make a preliminary guess at a possible probability function out of this? I still need ~300 more green buck hits and ~800 yellow buck hits, so I game to add more data, Though I'm taking a break for this evening.