There are MANY things we did wrong over the years. This is a list of those things we think you need to put some very careful thought and consideration into.

1. Set some goals - I always say, if you don't know where you are going, you won't know when you get there! Different rovers have different goals. Some want to just see how high of a microwave frequency they can effectively use. Some just want to get the highest possible score. Some want to see how long of a distance they can work. Some just want to have fun. Some want to give out as many "rare grids" as they can. Our goals in priority order are: 1 Have fun; 2 Have fun; 3 learn from mistakes and make improvements; 4 In the spirit of roving, go to as many grids as we can, as rare of grids as we can, and work as many stations as we can.
2. Operating efficiency - Time spent on thinking about operating efficiency is time well spent. Driving time is an essential, but the route and sites can be chosen so that there is not a lot of wasted time. Setup time just subtracts from operating time and should be minimized. Logging should be as simple as possible since it can detract from operating. We are using Roverlog and Tom Mayo has done an excellent job of improving it over the years to the point where it is very efficient and easy to use. He has been very cooperative in making changes that we and other rovers have made, and the current "product" is great. I would not consider any other logger for roving at this time. Operating should be simple and fast. Band switching should be easy and fast.
3. Site selection - Site selection is probably one of the hardest, most challenging, and time consuming tasks in the planning process. There are good starting points such as using topo maps (I tend to use Delorme Topo), navigational maps (I use both Delorme Street Atlas and Garmin Map Source), the N2MH site selector web site (http://www.n2mh.net/rovesite.htm), aerial photographs on the www, etc. Then after you pick some possibles, you almost HAVE to visit them to see about things like access ability (bad roads, chained gates, guard dogs, etc.), obstructions like trees and buildings, driving times, etc. Don't be afraid to solicit the advice of others on the potential sites. I have saved some driving by talking (E-Mailing) to people. You can spend more time driving to sites to check them out than you do driving during the contests! If the site is on private land, you need to find the owner and get permission. I am in the process of updating my web site to better describe our roving sites. Better photos will be added as we revisit sites.
4. Test Equipment - test equipment can end up costing A LOT! It seems to us to be a never ending set of compromises. If you want another station to hear you, two things are important. Your radiated power and your frequency. Fortunately the equipment to measure both can be had for reasonable cost. To measure power I use an old HP 432A power meter. Now that commercial organizations are replacing their old equipment with new digital stuff, the 432A can be had for a song. I got a couple of them off E-Bay for under $50 including thermistor mounts, although $100 is more typical. The thermistor mount can be harder to find than the meter itself, but watching for them can have good results. Some of the thermistor mounts may be off calibration, but many of them can be easily repaired. The meter in combination with reasonable good attenuators can effective and quite accurately measure rig/amplifier output power. This in combination with an appropriate selection of directional couplers to check return loss and you will have a good idea of your radiated power. Measuring frequency is not quite as cheap. I started with a small hand held Elenco counter which was under $100. It was fairly effective up to 1296, but at 2 and 3 GHz it was not stable or accurate enough. At 10GHz it wasn't nearly good enough. Since then I have gone to an HP Z3801 frequency standard and a reasonable Racal-Dana counter. I can now be quite accurate even at 10Ghz (within a couple hundred cycles at 10368.1) The biggest problem now is temperature drift of the local oscillators, but at least I can measure accurately. It would be nice to have good calibrated signal generators and a good spectrum analyzer, but unfortunately these still command a high price. Since the counter seems to stay very stable and accurate, I no longer take the Z3801 on the road roving. It has a VERY long warm up time.
5. Test Test Test - No amount of testing prior to the start of the contest is really wasted. In the few days before the contest I always measure power output, LO frequency, and antenna return loss of every band. An area I didn't check well enough before the Sept. 2005 contest was the deep cycle batteries, and I think one of them is getting a bit old. Test everything you can think of to test.
6. Site Dynamics - This is always an ongoing set of real work. We are very happy with our sites in FN32 FN33 and FN21. We are moderately happy with our sites in FN22, FN23 and FN31. We have yet to reach a good compromise in FN31 in terms of wasted travel time. Also we have thought about trying to get to some rarer grids like FN34 FN35, and/or maybe FN24 and FN25. The problem is a lot of extra driving, finding good sites Price of gas etc. However since FN34 and FN35 are very rare as compared to FN21 and FN31, we did try that route. We also have done roving in Florida as well as the skyline drive in VA and the DelMarVa.
7. Operation while mobile - We have had limited success in this area. We do manage to pick up some FM contacts from 6M through 440 as well as some SSB. Not a lot above that. If you look at our route, you will see we spend a lot of time driving through valleys with a lot of hills around. I would guess that if we drove down the Skyline Drive and Blue Ridge Parkway, we would have a lot more success while driving. We did have horizontal loops on the roof for 144 and 432 to use while mobile, but they seemed to have little value over just using the beams in whatever direction they were aimed. 6M is a lot more value (we do use the loop) if band conditions are OK. Currently we only operate 144 and 446 FM while driving
8. Operation with 2 ops - This is an area we have done a lot to improve. We did implement an earphone, PTT and Mic switching system for the two operator positions. This was a large improvement. We have had a bit of problems with RF getting into the Mic audio but did resolve that with better filtering. Also see our section on History and Evolution.