One technique is to compute takeoff roll based on wing loading (see AC90-89A), essentially a "most conservative" estimate and then adjust for density altitude effects and slope. A "best case" answer could be estimated by using test data from your airplane, if it's available, and then adjusting that. Britain's CAA has a good primer with some handy rules of thumb (Safety Sense Leaflet #7, Aircraft Performance) and the standard Koch Chart will provide estimates for altitude effects and correct for non-standard temperature. From a risk-management standpoint, it's best to compute "worst case."
RV-4 Estimate, 160 Horsepower Engine
Assumes 73% horsepower available at 9100' MSL under standard conditions
Standard temp at 9100 MSL: 26 degrees F
Worst Case Estimate, 1500 lbs gross takeoff weight, wing loading 13.6 lbs/foot: 850' required to accelerate and smoothly lift off at MSL (Per AC90-89A Minimum Recommended Runway Length Chart)
Density Altitude Adjustment:
Method 1, 10% correction factor per 1000' increase in DA: +774 feet
Method 2, Koch chart correction factor 180%: +680 feet
Net range for roll 1530-1624 feet corrected for density altitude affect (at standard temperature, no wind)
Slope Adjustment:
5% reduction per 2 degrees of slope: Net reduction 30%
Net range adjusted for slope: 1137-1043'
Surface adjustment: Dry grass (up to 8") on firm soil: Net increase 20%
Net range adjusted for altitude, slope, runway surface: 1252-1364 feet
So I'd guesstimate 1400' to get airborne. Not much margin for error, and no acellerate stop option, i.e., I'm committed to crash beyond the runway once I commit to takeoff.
Climb Performance Estimate:
Estimate rate of climb 992 FPM @ 102 MPH CAS (from bootstrap derived performance data corrected for density altitude affect). Koch chart estimates 70-75% reduction in rate of climb from standard, sea level conditions, so that works out to approximately 500 FPM, banding estimated climb performance between 500 FPM and 990 FPM.
Assuming worst case 500 FPM climb rate, climb gradient would be 345' per NM, so 3.4 degree climb angle required to clear obstacles.
Assuming best case 990 FPM climb rate, climb gradient would be 682' per nautical mile, or 6.8 degrees.
Therefore, I'd estimate climb gradient 3-6 degrees when looking at departure terrain.
If I had to adjust for non-standard temperature, I'd increase the distance 10% for each 10 degree C (18 degree F) increase. Headwind would increase performance, but I wouldn't factor it in (i.e., most conservative estimate), any tail wind would increase the estimate by 20% for each 5 knots of tailwind component.
Bottom line, I'd wait for my wingman to go first, and if he makes it, then I'd roll...
Cheers,
Vac