compass: small & light vs. robust

My current favorite is the Silva Guide (http://www.rei.com/product/408185)–a little lighter than my previous favorite–Silva Trekker.

Mirror does triple duty: (1) more accurate bearings (2) signal mirror (3) First Aid device–eye inspection tool (or other hard-to-view areas).

My usual compass is a Silva Ranger 27, a good compromise between something light and compact and fully featured. It's also much harder to lose than the "button" types often carried as emergency back-ups. In addition, if venturing to areas where a compass is absolutely essential, I'll usually have a button compass in my emergency kit, or else use the one on my match-case if I'm carrying that.

Here in the UK we are blessed with exceptionally good maps which, to some degree (no pun intended), allow a less accurate compass to be used adequately, but considering the weight of the Ranger 27 I don't think I'd bother taking something smaller, certainly not when on the moors where accuracy is vital due to the lack of obvious landmarks. If necessary I have a larger compass with various additional features, but for most of my walking this is just extra weight for no benefit.

When I lived in America I was surprised by the number of people who never carried a compass at all, presumably because the major trails are generally clearly marked. A different emphasis, perhaps: over here it's drummed into you early on that you don't set off without a compass, even on a day hike.

Something I can recommend is reading older books on navigation – Calvin Rutstrum wrote a good one – which very often have advice on navigating without a compass. These techniques can be very handy indeed and can give you extra confidence if using a very basic compass in difficult circumstances; and knowledge, as they say, weighs nothing.

I've recently ordered a Tru-Nord compensated compass, something a little different to the Silva, so I'll be interested to see how it compares.

I did a bunch of research on compass accuracy a couple of years ago in conjunction with the orienteering events I compete in, and training I was doing for the Boy Scouts.

I found a laboratory test that showed that a mirror-sighted compass was only more accurate if the mirror hinge was COMPLETELY parrellel. Otherwise, the angular error varied as function of how high the lid was raised. Their sample of Silva Ranger's did much better in this regard than their sample of Suunto mirrored compasses. Their sample of military lensatic compasses was MUCH more accurately put together and stayed accurate way longer.

Also, I found some US Army test results that said that sighting using the lensatic compass was much slower than the "waist hold" method, but NOT much more accurate "IF" the user locked their elbows into their waist, pivoted their whole body, and then just "looked up" to sight a distant object. The waist hold method has become the current recommendation in the US Army field manual for walking navigation (as oppossed to sighting a bearing for an artillery strike, etc).

Finally, I find from my orienteering experience that 1-2 degree accuracy versus 4-6 degree accuracy isn't that important… You'd think it was, from calculating how far you'd be off if you walked a mile an extra 3-4 degrees off to the side. However, in practice, it turns out that since you have to check the landmarks on the map to estimate distance anyway, you end up easily correcting for angular errors. Also, in my experience, angular error is mostly introduced by aiming at many intermediate objects when walking through forested areas that prevent you from sighting a truly distant object. ie: you have to go 3 miles on some heading, but can only see 150 feet, so you pick a tree and walk to it. Then you walk around the tree and sight the next tree, etc. It turns out when you walk around the tree so you can see /sight the next one, most people will systematically stand a little to the left (or right) even if they know about the problem and try not to (since the bearing isn't painted on the ground it's hard to know if you walked 180 degrees around the tree or 178 degrees). This causes them to introduce 1-2 degrees of drift at every re-sighting point. Periodically correcting for drift by using the landmarks on the map is almost automatic since you have to use it for estimate distance anyway (counting steps or using time doesn't work unless your stride length is constant).

I've also found, I only need a 3" baseplate to do map work (my longest bushwacking legs are usually 5-7 miles though). Theoretically, the longer baseplate should help aim better, but the "locked-in" waist hold method and correcting for error / drift using map landmarks makes the extra length unnecessary. So I carry a smaller compass (Silva 1070) with a 2"x3" baseplate since it fits in my pocket MUCH easier than my 2.5"x3.75" Suunto (3.75" is enough to get wedged sideways in my front pocket, and is bigger than the curve of my leg, so it makes the pant leg a little tighter). Also, the Silva's needle doesn't vibrate I'm standing there trying to take a bearing. The Suunto needle vibrates 1-2 degrees back and forth unless I focus on standing super still and controlling my breathing / heartrate, etc. I think it's because the Silva has thicker oil…

My compass doesn't have adjustable declination, but I just make a small tick mark on the bottom of the rotating inner module 11.5 degrees to the west (which is the declination in my area) with some nail polish and point the north needle at the tick mark rather than put the north needle in the "red shed" manufactured onto the face. If I travel across the country, I use nail polish remover and change the location of the mark. If I was doing a major east-west thru hike I would probably get one with adjustable declination since I often subtract wrong when I'm tired or hurried. For kids, putting the north needle in the shed is easier than using the tick mark. If you forget once, you're 11.5 degrees off course afterwards…