Terrain Movement Costs
Question: Is the TU usage for urban and rural items different? I see entries about broken urban fences, or Urban Garden Beds, but nothing for the rural equivalents.
The TU Usage is based on each tile's MCD data. The Farm terrain doesn't have fences or garden beds. It does however have a wheat field, which is in the list.
--Danial 15:52, 28 December 2005 (PST)
Was doing some testing earlier today on movement costs, and noted a few missing/differences. Firstly I guess there isnt anywhere I noticed on the page noting that diagonal movement costs 50% more (round down) than ortagonal moves.
Specific terrains - rough/heather stuff on agricultural maps is 5/7 I think, didnt notice it in the short list of 5 cost.
Skyranger ramp was a weird one in my testing, lateral moves were 8 as suggested in your list, but the move to the first step down coming out costs 4, the next step costs 12, then the last step 8 (plus 50% if moving diagonally again, so 18 if you had stepped out to the first layer then move diagonally across the ramp. Of course taking a dive of the side is always preferred - should just be able to tell the engineers to remove the ramp entirely because it makes it more difficult to get off the ship!
Moving between the ramp and gear of the Skyranger also cost 12, when you would expect 6 for basic terrain - if you are moving to hide beneath the front gear its quicker by 2 to go the outside route I believe (although more risky I guess).
I also got some oddities with raised terrain - the first layer seems to count as normal terrain, so you can walk over a one high hump, no matter how wide it is, at cost 4/6 all the way, its only when you move up to the second tier that the extra costs kick in and it starts costing 8/12. But then when you reach the whole next layer of the map, sometimes the top tier around that costs 4 again (but not always, couldnt work out the extra system).
Another item not noted is that turning costs 1 TU (edit: it is mentioned it has a fixed cost that I see, but nowhere lists that cost). But, if you move and it involves turns, you get any turns for free. So on standard terrain when facing orthagonally if you right click the square behind you, it costs 4 to turn. If you left click instead you will make the same 4 turns, then advance a square, and it still costs 4. This has saved my soldiers a number of times (as has generally knowing what each move will cost before I do it, when things get tight and you cant afford to be left holding a primed grenade with 1 less TU than it takes to throw, or standing adjacent to an armed alien with a stun rod that costs 1 or 2 TUs more than you now have remaining).
Would add this myself of course, but wondering because the data is apparently directly taken from the files so not sure if I have a different version or something?
--Sfnhltb 15:47, 25 February 2007 (PST)
Ramps are weird, yes. What's happening there is that the first step off the ramp can either be into "air" (same level, no extra cost) or onto the ramp (one level down, extra cost as stated).
Soldiers without flying suits end up in the same position: on the ramp. Soldiers in flying suits...end up in mid air, and have to waste 8 TUs going down a level before doing the standard "debark at side".
--Zaimoni 09:33, 26 February 2007 (CST)
Those all seem like valid comments to me, Sfnhltb. You're welcome to add to the page or fix any oversights. I think I fixed one or two. I'm not sure whether Danial took into account every possible combination of floor tile plus additional object (various plants, various broken things, etc.).
You've made some good points:
- Right, the diagonals are a glaring omission. Probly Dan (or whomever) wasn't thinking about something so basic.
- Good point about turning. It's actually an important tactical trick. All us vets probly long since figured it out, but the wiki's for basics, too.
- Thanks for the work on hills. I had also noticed that point about the lowest level being "free". Wouldn't hurt to add something.
Feel free to add this stuff to the page. I also just added a complete MCD database to the wiki, and put a link to the MCD page, on the TU page.
- MikeTheRed 09:02, 26 February 2007 (PST)
Experience Cap Probabilities
EthC, on second thought I'm thinking the chance of cap+1 for secondaries, is one third. Basically, three things can happen:
Cap-2: Roll 2, final value = cap Cap-1: Roll 1, final value = cap Cap-1: Roll 2, final value = cap+1
No other values or rolls matter in this regard, AFAIK. We assume we have approached the cap randomly, IOW, a player can't try to play with the rolls and influence them - but that's easy because the secondary stat would've been something much lower as a rookie, would've gotten lots of varying rolls as it increased, and then is constrained to 0-2 in the final leg of its increases. (The same cannot be said for primary-stat rolls because a person can use the experience counts to try to "edge" to cap-1, then make 11 actions to try for a 2-6 roll. Secondary stats are immune to manipulation. Not counting hacking, of course.)
This is not something I could've modelled easily, because a rigorous test would've required edging a lot of soldiers up from some secondary-stat score low enough that it would have been "random" by the time they approached the cap. Then rinse and repeat thousands of times. IOW, tons of real-world time.
I can say with certainty that if you are at cap-1, you are uniformly likely to roll 0, 1, or 2. Thus cap-1 acheives cap+1, one-third of the time. (Ok, 32.762% of the time, N=15,920.) But this is only an affirmation that XCOM uses the usual "roll within a range" (here, 0-2), not a true statistical proof of what happens when you approach the cap. AKA, it's one thing to hack a savegame so that all secondaries are at cap-1, then reload that a ton of times... that's 30 seconds per data dump, times the number of soldiers in the savegame. But it's way more work to do actual, real combats through multiple combats - and do this dozens or hundreds of different times.
Then again, it could be simulated simply by doing a zillion 0-2 rolls (start from 0, have a cap of 100, repeat a million times)... hmm.
Your thoughts, or anyone else's?
---MikeTheRed 20:10, 17 May 2006 (PDT)
Well, here's how I figured it:
First, I believe you have about a 50/50 chance of landing at Cap-1 or Cap-2 on the way to the cap. (As I understand it, there's a forensic accounting rule that says Cap-2, being a lower number, will be very slightly favored.)
From there, here are the possibilities:
- 50% chance of Cap
- 25% chance of Cap-1 followed by Cap
- 25% chance of Cap-1 followed by Cap+1
- 50% chance of Cap
- 50% chance of Cap+1
Averaging the 25% chance at Cap-2 and 50% at Cap-1 gives you 37.5% chance of hitting Cap+1.
Then again, I might be wrong.
--Ethereal Cereal 21:47, 17 May 2006 (PDT)
Heh. Okay, I was wrong. I ran the simulation you suggested: with 50,000 iterations, the result was unambiguous: cap 66.698% of the time, cap+1 33.302%. I'm not sure what your N=15,920 refers to, but that's not right either.
Now that I think about it, the 50% chance of being at Cap-1 already includes the possibility of getting there via Cap-2.
--Ethereal Cereal 23:32, 17 May 2006 (PDT)
Thanks for nailing that down, good work! Now we know for sure. The individual Secondary Stat sections and the Experience page should all be updated to say one third.
Nevermind the N=15,920. That's just how many times I sampled XCOM behavior for what secondary stats do when at cap-1 and a primary action was performed (via savegame reloads). But that's not the same thing as this "ultimate chance" of winding up with cap+1.
For what it's worth, if you've got the code set up, it might be easy to confirm my other chance, for getting a primary stat to cap+5 (Experience#Regarding_Caps). It should come out to a 5% chance in the event that a player starts doing 11+ actions when they get near the cap; in other words, just set your re-roller to roll 2-6 (instead of 0-2), start at 0, cap at 100, and see how often you get 105. This simulates a person who always maxxes their primaries and is not "paying attention" to just where they are relative to the cap.
Another possibility is for players to manipulate their primary actions and be sure they get to cap-1 (you can edge forward by 0-1 points by doing just 1-2 primaries) and then do 11+ actions to get a 2-6 roll when at cap-1... this gives you a 20% chance of cap+5. Of course that would take incredible dedication (and attention!), more than hardly anybody cares about, including myself. Still, it's interesting to know.
Most players probably shoot for 11+ actions on each combat but don't always get them (once they're rich and doing experience training), so for everything but psi skill, cap+5 is probably pretty rare... especially considering that probably few folks play a given game long enough to max anything besides psi skill. (Anybody correct me if wrong! We probably have a few very hardcore players here on the wiki, hehe.)
Ethereal, let me take this opportunity to thank you for all the good work you've been doing around the wiki. Some of the pages I worked on, like Experience Training and Profitability, were sorely in need of organization. Otherwise you've been doing lots of appropriate or useful edits here and there, plus have found two mistakes of mine. (What, me make a mistake? Hehe.) Thanks for being here! Maybe I'll get back into XCOM and the wiki in a big way some time again. There are some things I really want to test more. But not right now. :P
I'll try to get to that CMs table in Experience if I get the chance.
---MikeTheRed 14:35, 18 May 2006 (PDT)
Yep, cap+5 occurs 5% of the time when you're doing 11 actions each combat. Which seems less likely to me to happen in real life than some crazy hardcore player who does inch up to cap-1 then shoots for +5.
Thanks for the thanks. I've had the most fun doing Destroying Terrain and Reaction fire triggers, both of which revealed new combat strategies to me as I was researching them. I'm about to finish a rewrite of Research and after that I want to make the Main Page point to all the best info on the site -- right now it makes the site look like it's barely more than a reprint of the UFOPaedia.
The way I see it, Psionics is the next juicy research project -- it'd be nice to finally discover the correct formulas. A bunch of disarmed Ethereals trapped in the Avenger should make it easy to figure out, I think... I might take a swing at that one.
I think the site is still lacking a really comprehensive tactical guide, and a good guide on what your first $20mil should be spent on. We still need Zombie's data on all alien stats at all settings, and the alien and item pages need a bit more standardization. Other than that, everything I would ever want is here (on UFO Defense, anyhow).
(This has basically become an open letter to all ufopaedia editors :-)
--Ethereal Cereal 17:57, 18 May 2006 (PDT)
Sounds great, Ethereal. I just re-checked my breakout on cap+5 possibilities and now am wondering if it should be 1 in 21 (4.76%) instead of 1 in 20. Here are the possible outcomes:
# Start Add End =Cap+ 1 94 6 100 0 2 95 5 100 0 3 95 6 101 1 4 96 4 100 0 5 96 5 101 1 6 96 6 102 2 7 97 3 100 0 8 97 4 101 1 9 97 5 102 2 10 97 6 103 3 11 98 2 100 0 12 98 3 101 1 13 98 4 102 2 14 98 5 103 3 15 98 6 104 4 16 99 1 100 0 17 99 2 101 1 18 99 3 102 2 19 99 4 103 3 20 99 5 104 4 21 99 6 105 5
Where "#" is just a line counter. There are 21 possibilities (if I'm doing this right). The above can be summarized according to how often (Count) a particular value (Endpoint) is reached:
Endpoint Count Frequency 100 6 28.57% 101 5 23.81% 102 4 19.05% 103 3 14.29% 104 2 9.52% 105 1 4.76% ---- 21
If you get a chance, can you run your re-roller enough to distinguish between 5.00% and 4.76%?
Your "Destroying Terrain" page is nice... I found it a lot of fun to play with MCDs and tiles, eh? BombBloke's numeric tileset was a total blast. So to speak. In your discussion of Vulnerable Walls (or somewhere near there), you might put a mention or link to the oddness seen in most outer UFO walls. Only North and West walls (as defined in e.g. MapView) are true walls; the rest are objects... the difference being that objects allow HE to pass through them (which can definitely kill anything on the other side), whereas true walls have "directional blocking" and, e.g., with armor 100 stop any blaster HE from passing through (see pictures and surrounding text in Explosions#HE_Block).
Your "Triggers" page is very insightful... it does give ideas on how to play. Hmm... I thought my guys would reaction fire on aliens that were only turning in place (in a firing-squad situation). But I haven't played in months and could be wrong. It was a situation where aliens were blocked into a single alien-base room and couldn't get out even if they wanted to, so this may have caused them to do some "unusual turning" (in theory they're starting to move to get out, but then realize they can't, or something).
Which reminds me. At some point in the Experience page, you mentioned that you liked my point #3 re: reaction fire experience counters. There's more on that stuck in some discussion somewhere... in subsequent testing I found that the extra counts occured right when an alien died, in a situation where more than one of your soldiers could've reaction fired. Apparently all your soldiers's reactions are "queued" and the next guy's reaction counter is ticked before it checks whether the alien is still up. Thus, you have the potential to get extra reaction counts up to the number of aliens killed, although just who will get them (if you have a lot of shooters) is anybody's guess. Unless you peak at Unitref and see exactly who did get it. Let me go fix that up. Looking over my pages now in hindsight, there are quite a few places where it's a snapshot in time, and more was realized later. I should try to fix those because others don't know what came first and it looks conflicting.
Psionics is exactly what I want to investigate, too. If I get back to this. I may have found a way to automate the XCOM interface, which, if true, could mean tons of psi testing data. It can be done by hand, too. Whichever way, I figure it needs lots of data due to the variability involved.
Other stuff could definitely use an update. There was once a long, long Discussion for the Main Page. NKF did most of the most-recent reorganization and "graphics" for it. I had hoped he would reorganize and make e.g. Damage and Explosions in a much more obvious place, like under BS Combat. Looks like he didn't get around to it (nor did I).
Hey Zombie, post your alien stats already! FWIW, he verified for me that I had armor and health stats correct for my Kill Modelling. Last I heard months ago, he was "this" close to finalizing it.
Another thing I wish were fleshed out is the Country Funding under Economics. There's a fascinating but incredibly long thread on XCOMUFO that needs to be summarized there. Thank Zombie, as usual.
Okeydoke, enough typing for now :)
---MikeTheRed 18:44, 19 May 2006 (PDT)
The problem I have when working with the wiki is that whenever I look at the main page, I'm left dumbstruck and unsure of what to do. It's just that overwhelming.
I'd actually like to see the main menus for the three main games go into a sub sections of their own, so that we can get more creative with the index and list the highlights under each section. Everything's too squeezed together as it is.
200,000 iterations. I'm not copying the exact numbers because I'm lazy, but the result was crystal clear:
Endpoint Frequency 100 25% 101 25% 102 20% 103 15% 104 10% 105 5%
--Ethereal Cereal 20:13, 19 May 2006 (PDT)
Ye gods, I see my problem Eth - look how I put "99+1=100" in there. But the roll range is only 2-6. Huh. Snuck up on me. :) There should only be 20 possible values, with a count of 5 for 100 - just like you have.
Hiya NKF, long time no type! I seem to recall I made tons of suggestions - about what you might do, lol. We've all agreed that whoever gets around to doing any particular bit of work can make the call, if nobody else is having an opinion. Assuming they're competent (and based on what I've seen, that definitely includes you, Eth). I might rework some of the linkage under the main headings, but overall the main menu itself is fine for me. Eth, we (and esp. me) talked a lot about what the main headings should be, and I made a big point that the main page should not be an exhaustive dictionary of XCOM. It should be a place with a minimal number of headings (it used to be cluttered) that serves two audiences: newbies and the hardcore. All big/important pages should be within 1 or 2 links deep off the main page. (This does not count more esoteric pages, like my kill modelling.) In this way, newbies can quickly find overviews or details... and the hardcore can readily get to the nuance/reference pages they might want, like Damage, Reaction Triggers, whatever. In this view that I had, things like BS Overview or BS Combat would have things like Damage, Explosions, Triggers, Experience set in a more obvious place. Right now they are here and there, stuck under Mission Overviews or whatever.
Anyway. Despite the much talk, Eth, if you have an idea, go for it. You can bounce it off us here if you like (and if nobody responds, just go for it!). It's also possible to post prototypes here in Discussion, or make a separate prototype page for the Main Page (without making it be the Main Page yet).
The "Under The Hood" could use a good overhaul some day, too. I don't know how much you (Eth) go under the hood, but that grew organically... it should have a more direct link that opens onto all the game files and savegame files' links, not the odd breakout it currently has.
NKF, you said you'd like to see "the main menus for the three main games go into sub sections of their own". Do you mean have the Main Page be simply a link to the individual games, and those games would all have their own "main page"?
---MikeTheRed 20:41, 19 May 2006 (PDT)
While you have a 33% chance of breaking cap in a single combat with 79 TUs, shouldn't it be mentioned that over the course of several battles, you will break the cap 50% of the time, not 33%? Think about it. If you roll a 0, the next battle you'll roll again. And eventually, you'll roll either a 1 or a 2, thus giving you a 50% chance of overflowing cap. Arrow Quivershaft 18:31, 2 June 2007 (PDT)
- It's still a 33% chance of getting any of the three outcomes. You have 1/3 of a chance to get nothing and a re-roll, a 1/3 chance to get 1, and a 1/3 chance of getting 2 points. There is a 50% of getting either one of the two outcomes 2/3rds of the time though. It's still: (2/3) / (1/2) = 33.33..%. -NKF
- As long as you're clear on "ultimate result" (pretending rerolls are the next roll) and "this time only", yes: the 1/3 chance of reroll should be removed when estimating the ultimate result. However, it's a lot less confusing to focus on the "this time only", as that is what you see directly. -- Zaimoni 22:18 CDT June 3 2007